Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

merge latest, Jan 12 2024

Browse Source
This commit is contained in:
Hui Liu 2024-01-12 09:57:11 -08:00
commit 1e1e8be3a5
251 changed files with 9023 additions and 5012 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
.github/workflows/r_tests.yml vendored
View File

@ -25,7 +25,7 @@ jobs:
with:
submodules: 'true'
- uses: r-lib/actions/setup-r@11a22a908006c25fe054c4ef0ac0436b1de3edbe # v2.6.4
- uses: r-lib/actions/setup-r@e40ad904310fc92e96951c1b0d64f3de6cbe9e14 # v2.6.5
with:
r-version: ${{ matrix.config.r }}
@ -54,7 +54,7 @@ jobs:
matrix:
config:
- {os: windows-latest, r: 'release', compiler: 'mingw', build: 'autotools'}
- {os: windows-latest, r: '4.2.0', compiler: 'msvc', build: 'cmake'}
- {os: windows-latest, r: '4.3.0', compiler: 'msvc', build: 'cmake'}
env:
R_REMOTES_NO_ERRORS_FROM_WARNINGS: true
RSPM: ${{ matrix.config.rspm }}
@ -64,7 +64,7 @@ jobs:
with:
submodules: 'true'
- uses: r-lib/actions/setup-r@11a22a908006c25fe054c4ef0ac0436b1de3edbe # v2.6.4
- uses: r-lib/actions/setup-r@e40ad904310fc92e96951c1b0d64f3de6cbe9e14 # v2.6.5
with:
r-version: ${{ matrix.config.r }}

9
R-package/CMakeLists.txt
View File

@ -14,6 +14,15 @@ if(ENABLE_ALL_WARNINGS)
target_compile_options(xgboost-r PRIVATE -Wall -Wextra)
endif()
if(MSVC)
# https://github.com/microsoft/LightGBM/pull/6061
# MSVC doesn't work with anonymous types in structs. (R complex)
#
# syntax error: missing ';' before identifier 'private_data_c'
#
target_compile_definitions(xgboost-r PRIVATE -DR_LEGACY_RCOMPLEX)
endif()
target_compile_definitions(
xgboost-r PUBLIC
-DXGBOOST_STRICT_R_MODE=1

5
R-package/DESCRIPTION
View File

@ -58,12 +58,13 @@ Suggests:
float,
titanic
Depends:
R (>= 3.3.0)
R (>= 4.3.0)
Imports:
Matrix (>= 1.1-0),
methods,
data.table (>= 1.9.6),
jsonlite (>= 1.0),
jsonlite (>= 1.0)
Roxygen: list(markdown = TRUE)
RoxygenNote: 7.2.3
Encoding: UTF-8
SystemRequirements: GNU make, C++17

19
R-package/NAMESPACE
View File

@ -2,16 +2,19 @@
S3method("[",xgb.DMatrix)
S3method("dimnames<-",xgb.DMatrix)
S3method(coef,xgb.Booster)
S3method(dim,xgb.DMatrix)
S3method(dimnames,xgb.DMatrix)
S3method(getinfo,xgb.Booster)
S3method(getinfo,xgb.DMatrix)
S3method(predict,xgb.Booster)
S3method(predict,xgb.Booster.handle)
S3method(print,xgb.Booster)
S3method(print,xgb.DMatrix)
S3method(print,xgb.cv.synchronous)
S3method(setinfo,xgb.Booster)
S3method(setinfo,xgb.DMatrix)
S3method(slice,xgb.DMatrix)
S3method(variable.names,xgb.Booster)
export("xgb.attr<-")
export("xgb.attributes<-")
export("xgb.config<-")
@ -26,21 +29,27 @@ export(cb.save.model)
export(getinfo)
export(setinfo)
export(slice)
export(xgb.Booster.complete)
export(xgb.DMatrix)
export(xgb.DMatrix.hasinfo)
export(xgb.DMatrix.save)
export(xgb.attr)
export(xgb.attributes)
export(xgb.config)
export(xgb.copy.Booster)
export(xgb.create.features)
export(xgb.cv)
export(xgb.dump)
export(xgb.gblinear.history)
export(xgb.get.DMatrix.data)
export(xgb.get.DMatrix.num.non.missing)
export(xgb.get.DMatrix.qcut)
export(xgb.get.config)
export(xgb.get.num.boosted.rounds)
export(xgb.ggplot.deepness)
export(xgb.ggplot.importance)
export(xgb.ggplot.shap.summary)
export(xgb.importance)
export(xgb.is.same.Booster)
export(xgb.load)
export(xgb.load.raw)
export(xgb.model.dt.tree)
@ -52,13 +61,12 @@ export(xgb.plot.shap.summary)
export(xgb.plot.tree)
export(xgb.save)
export(xgb.save.raw)
export(xgb.serialize)
export(xgb.set.config)
export(xgb.train)
export(xgb.unserialize)
export(xgboost)
import(methods)
importClassesFrom(Matrix,dgCMatrix)
importClassesFrom(Matrix,dgRMatrix)
importClassesFrom(Matrix,dgeMatrix)
importFrom(Matrix,colSums)
importFrom(Matrix,sparse.model.matrix)
@ -82,8 +90,11 @@ importFrom(graphics,points)
importFrom(graphics,title)
importFrom(jsonlite,fromJSON)
importFrom(jsonlite,toJSON)
importFrom(methods,new)
importFrom(stats,coef)
importFrom(stats,median)
importFrom(stats,predict)
importFrom(stats,variable.names)
importFrom(utils,head)
importFrom(utils,object.size)
importFrom(utils,str)

84
R-package/R/callbacks.R
View File

@ -228,7 +228,7 @@ cb.reset.parameters <- function(new_params) {
})
if (!is.null(env$bst)) {
xgb.parameters(env$bst$handle) <- pars
xgb.parameters(env$bst) <- pars
} else {
for (fd in env$bst_folds)
xgb.parameters(fd$bst) <- pars
@ -333,13 +333,13 @@ cb.early.stop <- function(stopping_rounds, maximize = FALSE,
if (!is.null(env$bst)) {
if (!inherits(env$bst, 'xgb.Booster'))
stop("'bst' in the parent frame must be an 'xgb.Booster'")
if (!is.null(best_score <- xgb.attr(env$bst$handle, 'best_score'))) {
if (!is.null(best_score <- xgb.attr(env$bst, 'best_score'))) {
best_score <<- as.numeric(best_score)
best_iteration <<- as.numeric(xgb.attr(env$bst$handle, 'best_iteration')) + 1
best_msg <<- as.numeric(xgb.attr(env$bst$handle, 'best_msg'))
best_iteration <<- as.numeric(xgb.attr(env$bst, 'best_iteration')) + 1
best_msg <<- as.numeric(xgb.attr(env$bst, 'best_msg'))
} else {
xgb.attributes(env$bst$handle) <- list(best_iteration = best_iteration - 1,
best_score = best_score)
xgb.attributes(env$bst) <- list(best_iteration = best_iteration - 1,
best_score = best_score)
}
} else if (is.null(env$bst_folds) || is.null(env$basket)) {
stop("Parent frame has neither 'bst' nor ('bst_folds' and 'basket')")
@ -348,7 +348,7 @@ cb.early.stop <- function(stopping_rounds, maximize = FALSE,
finalizer <- function(env) {
if (!is.null(env$bst)) {
attr_best_score <- as.numeric(xgb.attr(env$bst$handle, 'best_score'))
attr_best_score <- as.numeric(xgb.attr(env$bst, 'best_score'))
if (best_score != attr_best_score) {
# If the difference is too big, throw an error
if (abs(best_score - attr_best_score) >= 1e-14) {
@ -358,9 +358,9 @@ cb.early.stop <- function(stopping_rounds, maximize = FALSE,
# If the difference is due to floating-point truncation, update best_score
best_score <- attr_best_score
}
env$bst$best_iteration <- best_iteration
env$bst$best_ntreelimit <- best_ntreelimit
env$bst$best_score <- best_score
xgb.attr(env$bst, "best_iteration") <- best_iteration
xgb.attr(env$bst, "best_ntreelimit") <- best_ntreelimit
xgb.attr(env$bst, "best_score") <- best_score
} else {
env$basket$best_iteration <- best_iteration
env$basket$best_ntreelimit <- best_ntreelimit
@ -412,11 +412,15 @@ cb.early.stop <- function(stopping_rounds, maximize = FALSE,
#' @param save_period save the model to disk after every
#' \code{save_period} iterations; 0 means save the model at the end.
#' @param save_name the name or path for the saved model file.
#'
#' Note that the format of the model being saved is determined by the file
#' extension specified here (see \link{xgb.save} for details about how it works).
#'
#' It can contain a \code{\link[base]{sprintf}} formatting specifier
#' to include the integer iteration number in the file name.
#' E.g., with \code{save_name} = 'xgboost_%04d.model',
#' the file saved at iteration 50 would be named "xgboost_0050.model".
#'
#' E.g., with \code{save_name} = 'xgboost_%04d.ubj',
#' the file saved at iteration 50 would be named "xgboost_0050.ubj".
#' @seealso \link{xgb.save}
#' @details
#' This callback function allows to save an xgb-model file, either periodically after each \code{save_period}'s or at the end.
#'
@ -430,7 +434,7 @@ cb.early.stop <- function(stopping_rounds, maximize = FALSE,
#' \code{\link{callbacks}}
#'
#' @export
cb.save.model <- function(save_period = 0, save_name = "xgboost.model") {
cb.save.model <- function(save_period = 0, save_name = "xgboost.ubj") {
if (save_period < 0)
stop("'save_period' cannot be negative")
@ -440,8 +444,13 @@ cb.save.model <- function(save_period = 0, save_name = "xgboost.model") {
stop("'save_model' callback requires the 'bst' booster object in its calling frame")
if ((save_period > 0 && (env$iteration - env$begin_iteration) %% save_period == 0) ||
(save_period == 0 && env$iteration == env$end_iteration))
xgb.save(env$bst, sprintf(save_name, env$iteration))
(save_period == 0 && env$iteration == env$end_iteration)) {
# Note: this throws a warning if the name doesn't have anything to format through 'sprintf'
suppressWarnings({
save_name <- sprintf(save_name, env$iteration)
})
xgb.save(env$bst, save_name)
}
}
attr(callback, 'call') <- match.call()
attr(callback, 'name') <- 'cb.save.model'
@ -512,8 +521,7 @@ cb.cv.predict <- function(save_models = FALSE) {
env$basket$pred <- pred
if (save_models) {
env$basket$models <- lapply(env$bst_folds, function(fd) {
xgb.attr(fd$bst, 'niter') <- env$end_iteration - 1
xgb.Booster.complete(xgb.handleToBooster(handle = fd$bst, raw = NULL), saveraw = TRUE)
return(fd$bst)
})
}
}
@ -665,7 +673,7 @@ cb.gblinear.history <- function(sparse = FALSE) {
} else { # xgb.cv:
cf <- vector("list", length(env$bst_folds))
for (i in seq_along(env$bst_folds)) {
dmp <- xgb.dump(xgb.handleToBooster(handle = env$bst_folds[[i]]$bst, raw = NULL))
dmp <- xgb.dump(env$bst_folds[[i]]$bst)
cf[[i]] <- as.numeric(grep('(booster|bias|weigh)', dmp, invert = TRUE, value = TRUE))
if (sparse) cf[[i]] <- as(cf[[i]], "sparseVector")
}
@ -685,14 +693,19 @@ cb.gblinear.history <- function(sparse = FALSE) {
callback
}
#' Extract gblinear coefficients history.
#'
#' A helper function to extract the matrix of linear coefficients' history
#' @title Extract gblinear coefficients history.
#' @description A helper function to extract the matrix of linear coefficients' history
#' from a gblinear model created while using the \code{cb.gblinear.history()}
#' callback.
#' @details Note that this is an R-specific function that relies on R attributes that
#' are not saved when using xgboost's own serialization functions like \link{xgb.load}
#' or \link{xgb.load.raw}.
#'
#' In order for a serialized model to be accepted by tgis function, one must use R
#' serializers such as \link{saveRDS}.
#' @param model either an \code{xgb.Booster} or a result of \code{xgb.cv()}, trained
#' using the \code{cb.gblinear.history()} callback.
#' using the \code{cb.gblinear.history()} callback, but \bold{not} a booster
#' loaded from \link{xgb.load} or \link{xgb.load.raw}.
#' @param class_index zero-based class index to extract the coefficients for only that
#' specific class in a multinomial multiclass model. When it is NULL, all the
#' coefficients are returned. Has no effect in non-multiclass models.
@ -713,20 +726,18 @@ xgb.gblinear.history <- function(model, class_index = NULL) {
stop("model must be an object of either xgb.Booster or xgb.cv.synchronous class")
is_cv <- inherits(model, "xgb.cv.synchronous")
if (is.null(model[["callbacks"]]) || is.null(model$callbacks[["cb.gblinear.history"]]))
if (is_cv) {
callbacks <- model$callbacks
} else {
callbacks <- attributes(model)$callbacks
}
if (is.null(callbacks) || is.null(callbacks$cb.gblinear.history))
stop("model must be trained while using the cb.gblinear.history() callback")
if (!is_cv) {
# extract num_class & num_feat from the internal model
dmp <- xgb.dump(model)
if (length(dmp) < 2 || dmp[2] != "bias:")
stop("It does not appear to be a gblinear model")
dmp <- dmp[-c(1, 2)]
n <- which(dmp == 'weight:')
if (length(n) != 1)
stop("It does not appear to be a gblinear model")
num_class <- n - 1
num_feat <- (length(dmp) - 4) / num_class
num_class <- xgb.num_class(model)
num_feat <- xgb.num_feature(model)
} else {
# in case of CV, the object is expected to have this info
if (model$params$booster != "gblinear")
@ -742,7 +753,7 @@ xgb.gblinear.history <- function(model, class_index = NULL) {
(class_index[1] < 0 || class_index[1] >= num_class))
stop("class_index has to be within [0,", num_class - 1, "]")
coef_path <- environment(model$callbacks$cb.gblinear.history)[["coefs"]]
coef_path <- environment(callbacks$cb.gblinear.history)[["coefs"]]
if (!is.null(class_index) && num_class > 1) {
coef_path <- if (is.list(coef_path)) {
lapply(coef_path,
@ -770,7 +781,8 @@ xgb.gblinear.history <- function(model, class_index = NULL) {
if (!is.null(eval_err)) {
if (length(eval_res) != length(eval_err))
stop('eval_res & eval_err lengths mismatch')
res <- paste0(sprintf("%s:%f+%f", enames, eval_res, eval_err), collapse = '\t')
# Note: UTF-8 code for plus/minus sign is U+00B1
res <- paste0(sprintf("%s:%f\U00B1%f", enames, eval_res, eval_err), collapse = '\t')
} else {
res <- paste0(sprintf("%s:%f", enames, eval_res), collapse = '\t')
}

130
R-package/R/utils.R
View File

@ -93,6 +93,14 @@ check.booster.params <- function(params, ...) {
interaction_constraints <- sapply(params[['interaction_constraints']], function(x) paste0('[', paste(x, collapse = ','), ']'))
params[['interaction_constraints']] <- paste0('[', paste(interaction_constraints, collapse = ','), ']')
}
# for evaluation metrics, should generate multiple entries per metric
if (NROW(params[['eval_metric']]) > 1) {
eval_metrics <- as.list(params[["eval_metric"]])
names(eval_metrics) <- rep("eval_metric", length(eval_metrics))
params_without_ev_metrics <- within(params, rm("eval_metric"))
params <- c(params_without_ev_metrics, eval_metrics)
}
return(params)
}
@ -140,19 +148,17 @@ check.custom.eval <- function(env = parent.frame()) {
# Update a booster handle for an iteration with dtrain data
xgb.iter.update <- function(booster_handle, dtrain, iter, obj) {
if (!identical(class(booster_handle), "xgb.Booster.handle")) {
stop("booster_handle must be of xgb.Booster.handle class")
}
xgb.iter.update <- function(bst, dtrain, iter, obj) {
if (!inherits(dtrain, "xgb.DMatrix")) {
stop("dtrain must be of xgb.DMatrix class")
}
handle <- xgb.get.handle(bst)
if (is.null(obj)) {
.Call(XGBoosterUpdateOneIter_R, booster_handle, as.integer(iter), dtrain)
.Call(XGBoosterUpdateOneIter_R, handle, as.integer(iter), dtrain)
} else {
pred <- predict(
booster_handle,
bst,
dtrain,
outputmargin = TRUE,
training = TRUE,
@ -160,23 +166,24 @@ xgb.iter.update <- function(booster_handle, dtrain, iter, obj) {
)
gpair <- obj(pred, dtrain)
n_samples <- dim(dtrain)[1]
grad <- gpair$grad
hess <- gpair$hess
msg <- paste(
"Since 2.1.0, the shape of the gradient and hessian is required to be ",
"(n_samples, n_targets) or (n_samples, n_classes).",
sep = ""
)
if (is.matrix(gpair$grad) && dim(gpair$grad)[1] != n_samples) {
warning(msg)
}
if (is.numeric(gpair$grad) && length(gpair$grad) != n_samples) {
warning(msg)
if ((is.matrix(grad) && dim(grad)[1] != n_samples) ||
(is.vector(grad) && length(grad) != n_samples) ||
(is.vector(grad) != is.vector(hess))) {
warning(paste(
"Since 2.1.0, the shape of the gradient and hessian is required to be ",
"(n_samples, n_targets) or (n_samples, n_classes). Will reshape assuming ",
"column-major order.",
sep = ""
))
grad <- matrix(grad, nrow = n_samples)
hess <- matrix(hess, nrow = n_samples)
}
gpair$grad <- matrix(gpair$grad, nrow = n_samples)
gpair$hess <- matrix(gpair$hess, nrow = n_samples)
.Call(
XGBoosterBoostOneIter_R, booster_handle, dtrain, iter, gpair$grad, gpair$hess
XGBoosterTrainOneIter_R, handle, dtrain, iter, grad, hess
)
}
return(TRUE)
@ -186,23 +193,22 @@ xgb.iter.update <- function(booster_handle, dtrain, iter, obj) {
# Evaluate one iteration.
# Returns a named vector of evaluation metrics
# with the names in a 'datasetname-metricname' format.
xgb.iter.eval <- function(booster_handle, watchlist, iter, feval) {
if (!identical(class(booster_handle), "xgb.Booster.handle"))
stop("class of booster_handle must be xgb.Booster.handle")
xgb.iter.eval <- function(bst, watchlist, iter, feval) {
handle <- xgb.get.handle(bst)
if (length(watchlist) == 0)
return(NULL)
evnames <- names(watchlist)
if (is.null(feval)) {
msg <- .Call(XGBoosterEvalOneIter_R, booster_handle, as.integer(iter), watchlist, as.list(evnames))
msg <- .Call(XGBoosterEvalOneIter_R, handle, as.integer(iter), watchlist, as.list(evnames))
mat <- matrix(strsplit(msg, '\\s+|:')[[1]][-1], nrow = 2)
res <- structure(as.numeric(mat[2, ]), names = mat[1, ])
} else {
res <- sapply(seq_along(watchlist), function(j) {
w <- watchlist[[j]]
## predict using all trees
preds <- predict(booster_handle, w, outputmargin = TRUE, iterationrange = c(1, 1))
preds <- predict(bst, w, outputmargin = TRUE, iterationrange = c(1, 1))
eval_res <- feval(preds, w)
out <- eval_res$value
names(out) <- paste0(evnames[j], "-", eval_res$metric)
@ -343,16 +349,45 @@ xgb.createFolds <- function(y, k) {
#' @name xgboost-deprecated
NULL
#' Do not use \code{\link[base]{saveRDS}} or \code{\link[base]{save}} for long-term archival of
#' models. Instead, use \code{\link{xgb.save}} or \code{\link{xgb.save.raw}}.
#' @title Model Serialization and Compatibility
#' @description
#'
#' It is a common practice to use the built-in \code{\link[base]{saveRDS}} function (or
#' \code{\link[base]{save}}) to persist R objects to the disk. While it is possible to persist
#' \code{xgb.Booster} objects using \code{\link[base]{saveRDS}}, it is not advisable to do so if
#' the model is to be accessed in the future. If you train a model with the current version of
#' XGBoost and persist it with \code{\link[base]{saveRDS}}, the model is not guaranteed to be
#' accessible in later releases of XGBoost. To ensure that your model can be accessed in future
#' releases of XGBoost, use \code{\link{xgb.save}} or \code{\link{xgb.save.raw}} instead.
#' When it comes to serializing XGBoost models, it's possible to use R serializers such as
#' \link{save} or \link{saveRDS} to serialize an XGBoost R model, but XGBoost also provides
#' its own serializers with better compatibility guarantees, which allow loading
#' said models in other language bindings of XGBoost.
#'
#' Note that an `xgb.Booster` object, outside of its core components, might also keep:\itemize{
#' \item Additional model configuration (accessible through \link{xgb.config}),
#' which includes model fitting parameters like `max_depth` and runtime parameters like `nthread`.
#' These are not necessarily useful for prediction/importance/plotting.
#' \item Additional R-specific attributes - e.g. results of callbacks, such as evaluation logs,
#' which are kept as a `data.table` object, accessible through `attributes(model)$evaluation_log`
#' if present.
#' }
#'
#' The first one (configurations) does not have the same compatibility guarantees as
#' the model itself, including attributes that are set and accessed through \link{xgb.attributes} - that is, such configuration
#' might be lost after loading the booster in a different XGBoost version, regardless of the
#' serializer that was used. These are saved when using \link{saveRDS}, but will be discarded
#' if loaded into an incompatible XGBoost version. They are not saved when using XGBoost's
#' serializers from its public interface including \link{xgb.save} and \link{xgb.save.raw}.
#'
#' The second ones (R attributes) are not part of the standard XGBoost model structure, and thus are
#' not saved when using XGBoost's own serializers. These attributes are only used for informational
#' purposes, such as keeping track of evaluation metrics as the model was fit, or saving the R
#' call that produced the model, but are otherwise not used for prediction / importance / plotting / etc.
#' These R attributes are only preserved when using R's serializers.
#'
#' Note that XGBoost models in R starting from version `2.1.0` and onwards, and XGBoost models
#' before version `2.1.0`; have a very different R object structure and are incompatible with
#' each other. Hence, models that were saved with R serializers live `saveRDS` or `save` before
#' version `2.1.0` will not work with latter `xgboost` versions and vice versa. Be aware that
#' the structure of R model objects could in theory change again in the future, so XGBoost's serializers
#' should be preferred for long-term storage.
#'
#' Furthermore, note that using the package `qs` for serialization will require version 0.26 or
#' higher of said package, and will have the same compatibility restrictions as R serializers.
#'
#' @details
#' Use \code{\link{xgb.save}} to save the XGBoost model as a stand-alone file. You may opt into
@ -365,26 +400,29 @@ NULL
#' The \code{\link{xgb.save.raw}} function is useful if you'd like to persist the XGBoost model
#' as part of another R object.
#'
#' Note: Do not use \code{\link{xgb.serialize}} to store models long-term. It persists not only the
#' model but also internal configurations and parameters, and its format is not stable across
#' multiple XGBoost versions. Use \code{\link{xgb.serialize}} only for checkpointing.
#' Use \link{saveRDS} if you require the R-specific attributes that a booster might have, such
#' as evaluation logs, but note that future compatibility of such objects is outside XGBoost's
#' control as it relies on R's serialization format (see e.g. the details section in
#' \link{serialize} and \link{save} from base R).
#'
#' For more details and explanation about model persistence and archival, consult the page
#' \url{https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html}.
#'
#' @examples
#' data(agaricus.train, package='xgboost')
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
#' eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
#' bst <- xgb.train(data = xgb.DMatrix(agaricus.train$data, label = agaricus.train$label),
#' max_depth = 2, eta = 1, nthread = 2, nrounds = 2,
#' objective = "binary:logistic")
#'
#' # Save as a stand-alone file; load it with xgb.load()
#' xgb.save(bst, 'xgb.model')
#' bst2 <- xgb.load('xgb.model')
#' fname <- file.path(tempdir(), "xgb_model.ubj")
#' xgb.save(bst, fname)
#' bst2 <- xgb.load(fname)
#'
#' # Save as a stand-alone file (JSON); load it with xgb.load()
#' xgb.save(bst, 'xgb.model.json')
#' bst2 <- xgb.load('xgb.model.json')
#' if (file.exists('xgb.model.json')) file.remove('xgb.model.json')
#' fname <- file.path(tempdir(), "xgb_model.json")
#' xgb.save(bst, fname)
#' bst2 <- xgb.load(fname)
#'
#' # Save as a raw byte vector; load it with xgb.load.raw()
#' xgb_bytes <- xgb.save.raw(bst)
@ -395,12 +433,12 @@ NULL
#' # Persist the R object. Here, saveRDS() is okay, since it doesn't persist
#' # xgb.Booster directly. What's being persisted is the future-proof byte representation
#' # as given by xgb.save.raw().
#' saveRDS(obj, 'my_object.rds')
#' fname <- file.path(tempdir(), "my_object.Rds")
#' saveRDS(obj, fname)
#' # Read back the R object
#' obj2 <- readRDS('my_object.rds')
#' obj2 <- readRDS(fname)
#' # Re-construct xgb.Booster object from the bytes
#' bst2 <- xgb.load.raw(obj2$xgb_model_bytes)
#' if (file.exists('my_object.rds')) file.remove('my_object.rds')
#'
#' @name a-compatibility-note-for-saveRDS-save
NULL

970
R-package/R/xgb.Booster.R
View File

File diff suppressed because it is too large Load Diff

451
R-package/R/xgb.DMatrix.R
View File

@ -8,13 +8,45 @@
#' a \code{dgRMatrix} object,
#' a \code{dsparseVector} object (only when making predictions from a fitted model, will be
#' interpreted as a row vector), or a character string representing a filename.
#' @param info a named list of additional information to store in the \code{xgb.DMatrix} object.
#' See \code{\link{setinfo}} for the specific allowed kinds of
#' @param label Label of the training data.
#' @param weight Weight for each instance.
#'
#' Note that, for ranking task, weights are per-group. In ranking task, one weight
#' is assigned to each group (not each data point). This is because we
#' only care about the relative ordering of data points within each group,
#' so it doesn't make sense to assign weights to individual data points.
#' @param base_margin Base margin used for boosting from existing model.
#'
#' In the case of multi-output models, one can also pass multi-dimensional base_margin.
#' @param missing a float value to represents missing values in data (used only when input is a dense matrix).
#' It is useful when a 0 or some other extreme value represents missing values in data.
#' @param silent whether to suppress printing an informational message after loading from a file.
#' @param feature_names Set names for features. Overrides column names in data
#' frame and matrix.
#' @param nthread Number of threads used for creating DMatrix.
#' @param ... the \code{info} data could be passed directly as parameters, without creating an \code{info} list.
#' @param group Group size for all ranking group.
#' @param qid Query ID for data samples, used for ranking.
#' @param label_lower_bound Lower bound for survival training.
#' @param label_upper_bound Upper bound for survival training.
#' @param feature_weights Set feature weights for column sampling.
#' @param enable_categorical Experimental support of specializing for categorical features.
#'
#' If passing 'TRUE' and 'data' is a data frame,
#' columns of categorical types will automatically
#' be set to be of categorical type (feature_type='c') in the resulting DMatrix.
#'
#' If passing 'FALSE' and 'data' is a data frame with categorical columns,
#' it will result in an error being thrown.
#'
#' If 'data' is not a data frame, this argument is ignored.
#'
#' JSON/UBJSON serialization format is required for this.
#'
#' @details
#' Note that DMatrix objects are not serializable through R functions such as \code{saveRDS} or \code{save}.
#' If a DMatrix gets serialized and then de-serialized (for example, when saving data in an R session or caching
#' chunks in an Rmd file), the resulting object will not be usable anymore and will need to be reconstructed
#' from the original source of data.
#'
#' @examples
#' data(agaricus.train, package='xgboost')
@ -24,21 +56,43 @@
#' dtrain <- with(
#' agaricus.train, xgb.DMatrix(data, label = label, nthread = nthread)
#' )
#' xgb.DMatrix.save(dtrain, 'xgb.DMatrix.data')
#' dtrain <- xgb.DMatrix('xgb.DMatrix.data')
#' if (file.exists('xgb.DMatrix.data')) file.remove('xgb.DMatrix.data')
#' fname <- file.path(tempdir(), "xgb.DMatrix.data")
#' xgb.DMatrix.save(dtrain, fname)
#' dtrain <- xgb.DMatrix(fname)
#' @export
xgb.DMatrix <- function(data, info = list(), missing = NA, silent = FALSE, nthread = NULL, ...) {
cnames <- NULL
xgb.DMatrix <- function(
data,
label = NULL,
weight = NULL,
base_margin = NULL,
missing = NA,
silent = FALSE,
feature_names = colnames(data),
nthread = NULL,
group = NULL,
qid = NULL,
label_lower_bound = NULL,
label_upper_bound = NULL,
feature_weights = NULL,
enable_categorical = FALSE
) {
if (!is.null(group) && !is.null(qid)) {
stop("Either one of 'group' or 'qid' should be NULL")
}
ctypes <- NULL
if (typeof(data) == "character") {
if (length(data) > 1)
stop("'data' has class 'character' and length ", length(data),
".\n 'data' accepts either a numeric matrix or a single filename.")
if (length(data) > 1) {
stop(
"'data' has class 'character' and length ", length(data),
".\n 'data' accepts either a numeric matrix or a single filename."
)
}
data <- path.expand(data)
handle <- .Call(XGDMatrixCreateFromFile_R, data, as.integer(silent))
} else if (is.matrix(data)) {
handle <- .Call(XGDMatrixCreateFromMat_R, data, missing, as.integer(NVL(nthread, -1)))
cnames <- colnames(data)
handle <- .Call(
XGDMatrixCreateFromMat_R, data, missing, as.integer(NVL(nthread, -1))
)
} else if (inherits(data, "dgCMatrix")) {
handle <- .Call(
XGDMatrixCreateFromCSC_R,
@ -49,7 +103,6 @@ xgb.DMatrix <- function(data, info = list(), missing = NA, silent = FALSE, nthre
missing,
as.integer(NVL(nthread, -1))
)
cnames <- colnames(data)
} else if (inherits(data, "dgRMatrix")) {
handle <- .Call(
XGDMatrixCreateFromCSR_R,
@ -60,7 +113,6 @@ xgb.DMatrix <- function(data, info = list(), missing = NA, silent = FALSE, nthre
missing,
as.integer(NVL(nthread, -1))
)
cnames <- colnames(data)
} else if (inherits(data, "dsparseVector")) {
indptr <- c(0L, as.integer(length(data@i)))
ind <- as.integer(data@i) - 1L
@ -73,23 +125,112 @@ xgb.DMatrix <- function(data, info = list(), missing = NA, silent = FALSE, nthre
missing,
as.integer(NVL(nthread, -1))
)
} else if (is.data.frame(data)) {
ctypes <- sapply(data, function(x) {
if (is.factor(x)) {
if (!enable_categorical) {
stop(
"When factor type is used, the parameter `enable_categorical`",
" must be set to TRUE."
)
}
"c"
} else if (is.integer(x)) {
"int"
} else if (is.logical(x)) {
"i"
} else {
if (!is.numeric(x)) {
stop("Invalid type in dataframe.")
}
"float"
}
})
## as.data.frame somehow converts integer/logical into real.
data <- as.data.frame(sapply(data, function(x) {
if (is.factor(x)) {
## XGBoost uses 0-based indexing.
as.numeric(x) - 1
} else {
x
}
}))
handle <- .Call(
XGDMatrixCreateFromDF_R, data, missing, as.integer(NVL(nthread, -1))
)
} else {
stop("xgb.DMatrix does not support construction from ", typeof(data))
}
dmat <- handle
attributes(dmat) <- list(class = "xgb.DMatrix")
if (!is.null(cnames)) {
setinfo(dmat, "feature_name", cnames)
attributes(dmat) <- list(
class = "xgb.DMatrix",
fields = new.env()
)
if (!is.null(label)) {
setinfo(dmat, "label", label)
}
if (!is.null(weight)) {
setinfo(dmat, "weight", weight)
}
if (!is.null(base_margin)) {
setinfo(dmat, "base_margin", base_margin)
}
if (!is.null(feature_names)) {
setinfo(dmat, "feature_name", feature_names)
}
if (!is.null(group)) {
setinfo(dmat, "group", group)
}
if (!is.null(qid)) {
setinfo(dmat, "qid", qid)
}
if (!is.null(label_lower_bound)) {
setinfo(dmat, "label_lower_bound", label_lower_bound)
}
if (!is.null(label_upper_bound)) {
setinfo(dmat, "label_upper_bound", label_upper_bound)
}
if (!is.null(feature_weights)) {
setinfo(dmat, "feature_weights", feature_weights)
}
if (!is.null(ctypes)) {
setinfo(dmat, "feature_type", ctypes)
}
info <- append(info, list(...))
for (i in seq_along(info)) {
p <- info[i]
setinfo(dmat, names(p), p[[1]])
}
return(dmat)
}
#' @title Check whether DMatrix object has a field
#' @description Checks whether an xgb.DMatrix object has a given field assigned to
#' it, such as weights, labels, etc.
#' @param object The DMatrix object to check for the given \code{info} field.
#' @param info The field to check for presence or absence in \code{object}.
#' @seealso \link{xgb.DMatrix}, \link{getinfo.xgb.DMatrix}, \link{setinfo.xgb.DMatrix}
#' @examples
#' library(xgboost)
#' x <- matrix(1:10, nrow = 5)
#' dm <- xgb.DMatrix(x, nthread = 1)
#'
#' # 'dm' so far doesn't have any fields set
#' xgb.DMatrix.hasinfo(dm, "label")
#'
#' # Fields can be added after construction
#' setinfo(dm, "label", 1:5)
#' xgb.DMatrix.hasinfo(dm, "label")
#' @export
xgb.DMatrix.hasinfo <- function(object, info) {
if (!inherits(object, "xgb.DMatrix")) {
stop("Object is not an 'xgb.DMatrix'.")
}
if (.Call(XGCheckNullPtr_R, object)) {
warning("xgb.DMatrix object is invalid. Must be constructed again.")
return(FALSE)
}
return(NVL(attr(object, "fields")[[info]], FALSE))
}
# get dmatrix from data, label
# internal helper method
@ -194,26 +335,38 @@ dimnames.xgb.DMatrix <- function(x) {
}
#' Get information of an xgb.DMatrix object
#'
#' Get information of an xgb.DMatrix object
#' @param object Object of class \code{xgb.DMatrix}
#' @title Get or set information of xgb.DMatrix and xgb.Booster objects
#' @param object Object of class \code{xgb.DMatrix} of `xgb.Booster`.
#' @param name the name of the information field to get (see details)
#' @param ... other parameters
#'
#' @return For `getinfo`, will return the requested field. For `setinfo`, will always return value `TRUE`
#' if it succeeds.
#' @details
#' The \code{name} field can be one of the following:
#' The \code{name} field can be one of the following for `xgb.DMatrix`:
#'
#' \itemize{
#' \item \code{label}: label XGBoost learn from ;
#' \item \code{weight}: to do a weight rescale ;
#' \item \code{base_margin}: base margin is the base prediction XGBoost will boost from ;
#' \item \code{nrow}: number of rows of the \code{xgb.DMatrix}.
#' \item \code{label}
#' \item \code{weight}
#' \item \code{base_margin}
#' \item \code{label_lower_bound}
#' \item \code{label_upper_bound}
#' \item \code{group}
#' \item \code{feature_type}
#' \item \code{feature_name}
#' \item \code{nrow}
#' }
#' See the documentation for \link{xgb.DMatrix} for more information about these fields.
#'
#' For `xgb.Booster`, can be one of the following:
#' \itemize{
#' \item \code{feature_type}
#' \item \code{feature_name}
#' }
#'
#' \code{group} can be setup by \code{setinfo} but can't be retrieved by \code{getinfo}.
#' Note that, while 'qid' cannot be retrieved, it's possible to get the equivalent 'group'
#' for a DMatrix that had 'qid' assigned.
#'
#' \bold{Important}: when calling `setinfo`, the objects are modified in-place. See
#' \link{xgb.copy.Booster} for an idea of this in-place assignment works.
#' @examples
#' data(agaricus.train, package='xgboost')
#' dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
@ -225,49 +378,60 @@ dimnames.xgb.DMatrix <- function(x) {
#' stopifnot(all(labels2 == 1-labels))
#' @rdname getinfo
#' @export
getinfo <- function(object, ...) UseMethod("getinfo")
getinfo <- function(object, name) UseMethod("getinfo")
#' @rdname getinfo
#' @export
getinfo.xgb.DMatrix <- function(object, name, ...) {
getinfo.xgb.DMatrix <- function(object, name) {
allowed_int_fields <- 'group'
allowed_float_fields <- c(
'label', 'weight', 'base_margin',
'label_lower_bound', 'label_upper_bound'
)
allowed_str_fields <- c("feature_type", "feature_name")
allowed_fields <- c(allowed_float_fields, allowed_int_fields, allowed_str_fields, 'nrow')
if (typeof(name) != "character" ||
length(name) != 1 ||
!name %in% c('label', 'weight', 'base_margin', 'nrow',
'label_lower_bound', 'label_upper_bound', "feature_type", "feature_name")) {
stop(
"getinfo: name must be one of the following\n",
" 'label', 'weight', 'base_margin', 'nrow', 'label_lower_bound', 'label_upper_bound', 'feature_type', 'feature_name'"
)
!name %in% allowed_fields) {
stop("getinfo: name must be one of the following\n",
paste(paste0("'", allowed_fields, "'"), collapse = ", "))
}
if (name == "feature_name" || name == "feature_type") {
ret <- .Call(XGDMatrixGetStrFeatureInfo_R, object, name)
} else if (name != "nrow") {
ret <- .Call(XGDMatrixGetInfo_R, object, name)
} else {
if (name == "nrow") {
ret <- nrow(object)
} else if (name %in% allowed_str_fields) {
ret <- .Call(XGDMatrixGetStrFeatureInfo_R, object, name)
} else if (name %in% allowed_float_fields) {
ret <- .Call(XGDMatrixGetFloatInfo_R, object, name)
if (length(ret) > nrow(object)) {
ret <- matrix(ret, nrow = nrow(object), byrow = TRUE)
}
} else if (name %in% allowed_int_fields) {
if (name == "group") {
name <- "group_ptr"
}
ret <- .Call(XGDMatrixGetUIntInfo_R, object, name)
if (length(ret) > nrow(object)) {
ret <- matrix(ret, nrow = nrow(object), byrow = TRUE)
}
}
if (length(ret) == 0) return(NULL)
return(ret)
}
#' Set information of an xgb.DMatrix object
#'
#' Set information of an xgb.DMatrix object
#'
#' @param object Object of class "xgb.DMatrix"
#' @param name the name of the field to get
#' @rdname getinfo
#' @param info the specific field of information to set
#' @param ... other parameters
#'
#' @details
#' The \code{name} field can be one of the following:
#' See the documentation for \link{xgb.DMatrix} for possible fields that can be set
#' (which correspond to arguments in that function).
#'
#' \itemize{
#' \item \code{label}: label XGBoost learn from ;
#' \item \code{weight}: to do a weight rescale ;
#' \item \code{base_margin}: base margin is the base prediction XGBoost will boost from ;
#' \item \code{group}: number of rows in each group (to use with \code{rank:pairwise} objective).
#' Note that the following fields are allowed in the construction of an \code{xgb.DMatrix}
#' but \bold{aren't} allowed here:\itemize{
#' \item data
#' \item missing
#' \item silent
#' \item nthread
#' }
#'
#' @examples
@ -278,52 +442,61 @@ getinfo.xgb.DMatrix <- function(object, name, ...) {
#' setinfo(dtrain, 'label', 1-labels)
#' labels2 <- getinfo(dtrain, 'label')
#' stopifnot(all.equal(labels2, 1-labels))
#' @rdname setinfo
#' @export
setinfo <- function(object, ...) UseMethod("setinfo")
setinfo <- function(object, name, info) UseMethod("setinfo")
#' @rdname setinfo
#' @rdname getinfo
#' @export
setinfo.xgb.DMatrix <- function(object, name, info, ...) {
setinfo.xgb.DMatrix <- function(object, name, info) {
.internal.setinfo.xgb.DMatrix(object, name, info)
attr(object, "fields")[[name]] <- TRUE
return(TRUE)
}
.internal.setinfo.xgb.DMatrix <- function(object, name, info) {
if (name == "label") {
if (length(info) != nrow(object))
if (NROW(info) != nrow(object))
stop("The length of labels must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "label_lower_bound") {
if (length(info) != nrow(object))
if (NROW(info) != nrow(object))
stop("The length of lower-bound labels must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "label_upper_bound") {
if (length(info) != nrow(object))
if (NROW(info) != nrow(object))
stop("The length of upper-bound labels must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "weight") {
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "base_margin") {
# if (length(info)!=nrow(object))
# stop("The length of base margin must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "group") {
if (sum(info) != nrow(object))
stop("The sum of groups must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, as.integer(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "qid") {
if (NROW(info) != nrow(object))
stop("The length of qid assignments must equal to the number of rows in the input data")
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
if (name == "feature_weights") {
if (length(info) != ncol(object)) {
if (NROW(info) != ncol(object)) {
stop("The number of feature weights must equal to the number of columns in the input data")
}
.Call(XGDMatrixSetInfo_R, object, name, as.numeric(info))
.Call(XGDMatrixSetInfo_R, object, name, info)
return(TRUE)
}
@ -353,6 +526,111 @@ setinfo.xgb.DMatrix <- function(object, name, info, ...) {
stop("setinfo: unknown info name ", name)
}
#' @title Get Quantile Cuts from DMatrix
#' @description Get the quantile cuts (a.k.a. borders) from an `xgb.DMatrix`
#' that has been quantized for the histogram method (`tree_method="hist"`).
#'
#' These cuts are used in order to assign observations to bins - i.e. these are ordered
#' boundaries which are used to determine assignment condition `border_low < x < border_high`.
#' As such, the first and last bin will be outside of the range of the data, so as to include
#' all of the observations there.
#'
#' If a given column has 'n' bins, then there will be 'n+1' cuts / borders for that column,
#' which will be output in sorted order from lowest to highest.
#'
#' Different columns can have different numbers of bins according to their range.
#' @param dmat An `xgb.DMatrix` object, as returned by \link{xgb.DMatrix}.
#' @param output Output format for the quantile cuts. Possible options are:\itemize{
#' \item `"list"` will return the output as a list with one entry per column, where
#' each column will have a numeric vector with the cuts. The list will be named if
#' `dmat` has column names assigned to it.
#' \item `"arrays"` will return a list with entries `indptr` (base-0 indexing) and
#' `data`. Here, the cuts for column 'i' are obtained by slicing 'data' from entries
#' `indptr[i]+1` to `indptr[i+1]`.
#' }
#' @return The quantile cuts, in the format specified by parameter `output`.
#' @examples
#' library(xgboost)
#' data(mtcars)
#' y <- mtcars$mpg
#' x <- as.matrix(mtcars[, -1])
#' dm <- xgb.DMatrix(x, label = y, nthread = 1)
#'
#' # DMatrix is not quantized right away, but will be once a hist model is generated
#' model <- xgb.train(
#' data = dm,
#' params = list(
#' tree_method = "hist",
#' max_bin = 8,
#' nthread = 1
#' ),
#' nrounds = 3
#' )
#'
#' # Now can get the quantile cuts
#' xgb.get.DMatrix.qcut(dm)
#' @export
xgb.get.DMatrix.qcut <- function(dmat, output = c("list", "arrays")) { # nolint
stopifnot(inherits(dmat, "xgb.DMatrix"))
output <- head(output, 1L)
stopifnot(output %in% c("list", "arrays"))
res <- .Call(XGDMatrixGetQuantileCut_R, dmat)
if (output == "arrays") {
return(res)
} else {
feature_names <- getinfo(dmat, "feature_name")
ncols <- length(res$indptr) - 1
out <- lapply(
seq(1, ncols),
function(col) {
st <- res$indptr[col]
end <- res$indptr[col + 1]
if (end <= st) {
return(numeric())
}
return(res$data[seq(1 + st, end)])
}
)
if (NROW(feature_names)) {
names(out) <- feature_names
}
return(out)
}
}
#' @title Get Number of Non-Missing Entries in DMatrix
#' @param dmat An `xgb.DMatrix` object, as returned by \link{xgb.DMatrix}.
#' @return The number of non-missing entries in the DMatrix
#' @export
xgb.get.DMatrix.num.non.missing <- function(dmat) { # nolint
stopifnot(inherits(dmat, "xgb.DMatrix"))
return(.Call(XGDMatrixNumNonMissing_R, dmat))
}
#' @title Get DMatrix Data
#' @param dmat An `xgb.DMatrix` object, as returned by \link{xgb.DMatrix}.
#' @return The data held in the DMatrix, as a sparse CSR matrix (class `dgRMatrix`
#' from package `Matrix`). If it had feature names, these will be added as column names
#' in the output.
#' @export
xgb.get.DMatrix.data <- function(dmat) {
stopifnot(inherits(dmat, "xgb.DMatrix"))
res <- .Call(XGDMatrixGetDataAsCSR_R, dmat)
out <- methods::new("dgRMatrix")
nrows <- as.integer(length(res$indptr) - 1)
out@p <- res$indptr
out@j <- res$indices
out@x <- res$data
out@Dim <- as.integer(c(nrows, res$ncols))
feature_names <- getinfo(dmat, "feature_name")
dim_names <- list(NULL, NULL)
if (NROW(feature_names)) {
dim_names[[2L]] <- feature_names
}
out@Dimnames <- dim_names
return(out)
}
#' Get a new DMatrix containing the specified rows of
#' original xgb.DMatrix object
@ -363,7 +641,6 @@ setinfo.xgb.DMatrix <- function(object, name, info, ...) {
#' @param object Object of class "xgb.DMatrix"
#' @param idxset a integer vector of indices of rows needed
#' @param colset currently not used (columns subsetting is not available)
#' @param ... other parameters (currently not used)
#'
#' @examples
#' data(agaricus.train, package='xgboost')
@ -377,11 +654,11 @@ setinfo.xgb.DMatrix <- function(object, name, info, ...) {
#'
#' @rdname slice.xgb.DMatrix
#' @export
slice <- function(object, ...) UseMethod("slice")
slice <- function(object, idxset) UseMethod("slice")
#' @rdname slice.xgb.DMatrix
#' @export
slice.xgb.DMatrix <- function(object, idxset, ...) {
slice.xgb.DMatrix <- function(object, idxset) {
if (!inherits(object, "xgb.DMatrix")) {
stop("object must be xgb.DMatrix")
}
@ -431,11 +708,15 @@ slice.xgb.DMatrix <- function(object, idxset, ...) {
#' @method print xgb.DMatrix
#' @export
print.xgb.DMatrix <- function(x, verbose = FALSE, ...) {
if (.Call(XGCheckNullPtr_R, x)) {
cat("INVALID xgb.DMatrix object. Must be constructed anew.\n")
return(invisible(x))
}
cat('xgb.DMatrix dim:', nrow(x), 'x', ncol(x), ' info: ')
infos <- character(0)
if (length(getinfo(x, 'label')) > 0) infos <- 'label'
if (length(getinfo(x, 'weight')) > 0) infos <- c(infos, 'weight')
if (length(getinfo(x, 'base_margin')) > 0) infos <- c(infos, 'base_margin')
if (xgb.DMatrix.hasinfo(x, 'label')) infos <- 'label'
if (xgb.DMatrix.hasinfo(x, 'weight')) infos <- c(infos, 'weight')
if (xgb.DMatrix.hasinfo(x, 'base_margin')) infos <- c(infos, 'base_margin')
if (length(infos) == 0) infos <- 'NA'
cat(infos)
cnames <- colnames(x)

6
R-package/R/xgb.DMatrix.save.R
View File

@ -8,9 +8,9 @@
#' @examples
#' data(agaricus.train, package='xgboost')
#' dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
#' xgb.DMatrix.save(dtrain, 'xgb.DMatrix.data')
#' dtrain <- xgb.DMatrix('xgb.DMatrix.data')
#' if (file.exists('xgb.DMatrix.data')) file.remove('xgb.DMatrix.data')
#' fname <- file.path(tempdir(), "xgb.DMatrix.data")
#' xgb.DMatrix.save(dtrain, fname)
#' dtrain <- xgb.DMatrix(fname)
#' @export
xgb.DMatrix.save <- function(dmatrix, fname) {
if (typeof(fname) != "character")

2
R-package/R/xgb.create.features.R
View File

@ -51,7 +51,7 @@
#' dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
#' dtest <- with(agaricus.test, xgb.DMatrix(data, label = label, nthread = 2))
#'
#' param <- list(max_depth=2, eta=1, silent=1, objective='binary:logistic')
#' param <- list(max_depth=2, eta=1, objective='binary:logistic')
#' nrounds = 4
#'
#' bst = xgb.train(params = param, data = dtrain, nrounds = nrounds, nthread = 2)

24
R-package/R/xgb.cv.R
View File

@ -126,6 +126,9 @@ xgb.cv <- function(params = list(), data, nrounds, nfold, label = NULL, missing
early_stopping_rounds = NULL, maximize = NULL, callbacks = list(), ...) {
check.deprecation(...)
if (inherits(data, "xgb.DMatrix") && .Call(XGCheckNullPtr_R, data)) {
stop("'data' is an invalid 'xgb.DMatrix' object. Must be constructed again.")
}
params <- check.booster.params(params, ...)
# TODO: should we deprecate the redundant 'metrics' parameter?
@ -136,7 +139,7 @@ xgb.cv <- function(params = list(), data, nrounds, nfold, label = NULL, missing
check.custom.eval()
# Check the labels
if ((inherits(data, 'xgb.DMatrix') && is.null(getinfo(data, 'label'))) ||
if ((inherits(data, 'xgb.DMatrix') && !xgb.DMatrix.hasinfo(data, 'label')) ||
(!inherits(data, 'xgb.DMatrix') && is.null(label))) {
stop("Labels must be provided for CV either through xgb.DMatrix, or through 'label=' when 'data' is matrix")
} else if (inherits(data, 'xgb.DMatrix')) {
@ -201,13 +204,13 @@ xgb.cv <- function(params = list(), data, nrounds, nfold, label = NULL, missing
dtrain <- slice(dall, unlist(folds[-k]))
else
dtrain <- slice(dall, train_folds[[k]])
handle <- xgb.Booster.handle(
bst <- xgb.Booster(
params = params,
cachelist = list(dtrain, dtest),
modelfile = NULL,
handle = NULL
modelfile = NULL
)
list(dtrain = dtrain, bst = handle, watchlist = list(train = dtrain, test = dtest), index = folds[[k]])
bst <- bst$bst
list(dtrain = dtrain, bst = bst, watchlist = list(train = dtrain, test = dtest), index = folds[[k]])
})
rm(dall)
# a "basket" to collect some results from callbacks
@ -228,21 +231,22 @@ xgb.cv <- function(params = list(), data, nrounds, nfold, label = NULL, missing
msg <- lapply(bst_folds, function(fd) {
xgb.iter.update(
booster_handle = fd$bst,
bst = fd$bst,
dtrain = fd$dtrain,
iter = iteration - 1,
obj = obj
)
xgb.iter.eval(
booster_handle = fd$bst,
bst = fd$bst,
watchlist = fd$watchlist,
iter = iteration - 1,
feval = feval
)
})
msg <- simplify2array(msg)
bst_evaluation <- rowMeans(msg)
bst_evaluation_err <- sqrt(rowMeans(msg^2) - bst_evaluation^2) # nolint
# Note: these variables might look unused here, but they are used in the callbacks
bst_evaluation <- rowMeans(msg) # nolint
bst_evaluation_err <- apply(msg, 1, sd) # nolint
for (f in cb$post_iter) f()
@ -263,7 +267,7 @@ xgb.cv <- function(params = list(), data, nrounds, nfold, label = NULL, missing
ret <- c(ret, basket)
class(ret) <- 'xgb.cv.synchronous'
invisible(ret)
return(invisible(ret))
}

24
R-package/R/xgb.dump.R
View File

@ -13,7 +13,10 @@
#' When this option is on, the model dump contains two additional values:
#' gain is the approximate loss function gain we get in each split;
#' cover is the sum of second order gradient in each node.
#' @param dump_format either 'text' or 'json' format could be specified.
#' @param dump_format either 'text', 'json', or 'dot' (graphviz) format could be specified.
#'
#' Format 'dot' for a single tree can be passed directly to packages that consume this format
#' for graph visualization, such as function [DiagrammeR::grViz()]
#' @param ... currently not used
#'
#' @return
@ -37,9 +40,13 @@
#' # print in JSON format:
#' cat(xgb.dump(bst, with_stats = TRUE, dump_format='json'))
#'
#' # plot first tree leveraging the 'dot' format
#' if (requireNamespace('DiagrammeR', quietly = TRUE)) {
#' DiagrammeR::grViz(xgb.dump(bst, dump_format = "dot")[[1L]])
#' }
#' @export
xgb.dump <- function(model, fname = NULL, fmap = "", with_stats = FALSE,
dump_format = c("text", "json"), ...) {
dump_format = c("text", "json", "dot"), ...) {
check.deprecation(...)
dump_format <- match.arg(dump_format)
if (!inherits(model, "xgb.Booster"))
@ -49,9 +56,16 @@ xgb.dump <- function(model, fname = NULL, fmap = "", with_stats = FALSE,
if (!(is.null(fmap) || is.character(fmap)))
stop("fmap: argument must be a character string (when provided)")
model <- xgb.Booster.complete(model)
model_dump <- .Call(XGBoosterDumpModel_R, model$handle, NVL(fmap, "")[1], as.integer(with_stats),
as.character(dump_format))
model_dump <- .Call(
XGBoosterDumpModel_R,
xgb.get.handle(model),
NVL(fmap, "")[1],
as.integer(with_stats),
as.character(dump_format)
)
if (dump_format == "dot") {
return(sapply(model_dump, function(x) gsub("^booster\\[\\d+\\]\\n", "\\1", x)))
}
if (is.null(fname))
model_dump <- gsub('\t', '', model_dump, fixed = TRUE)

31
R-package/R/xgb.ggplot.R
View File

@ -127,22 +127,20 @@ xgb.ggplot.shap.summary <- function(data, shap_contrib = NULL, features = NULL,
p
}
#' Combine and melt feature values and SHAP contributions for sample
#' observations.
#' Combine feature values and SHAP values
#'
#' Conforms to data format required for ggplot functions.
#' Internal function used to combine and melt feature values and SHAP contributions
#' as required for ggplot functions related to SHAP.
#'
#' Internal utility function.
#' @param data_list The result of `xgb.shap.data()`.
#' @param normalize Whether to standardize feature values to mean 0 and
#' standard deviation 1. This is useful for comparing multiple features on the same
#' plot. Default is \code{FALSE}.
#'
#' @param data_list List containing 'data' and 'shap_contrib' returned by
#' \code{xgb.shap.data()}.
#' @param normalize Whether to standardize feature values to have mean 0 and
#' standard deviation 1 (useful for comparing multiple features on the same
#' plot). Default \code{FALSE}.
#'
#' @return A data.table containing the observation ID, the feature name, the
#' @return A `data.table` containing the observation ID, the feature name, the
#' feature value (normalized if specified), and the SHAP contribution value.
#' @noRd
#' @keywords internal
prepare.ggplot.shap.data <- function(data_list, normalize = FALSE) {
data <- data_list[["data"]]
shap_contrib <- data_list[["shap_contrib"]]
@ -163,15 +161,16 @@ prepare.ggplot.shap.data <- function(data_list, normalize = FALSE) {
p_data
}
#' Scale feature value to have mean 0, standard deviation 1
#' Scale feature values
#'
#' This is used to compare multiple features on the same plot.
#' Internal utility function
#' Internal function that scales feature values to mean 0 and standard deviation 1.
#' Useful to compare multiple features on the same plot.
#'
#' @param x Numeric vector
#' @param x Numeric vector.
#'
#' @return Numeric vector with mean 0 and sd 1.
#' @return Numeric vector with mean 0 and standard deviation 1.
#' @noRd
#' @keywords internal
normalize <- function(x) {
loc <- mean(x, na.rm = TRUE)
scale <- stats::sd(x, na.rm = TRUE)

162
R-package/R/xgb.importance.R
View File

@ -1,83 +1,115 @@
#' Importance of features in a model.
#' Feature importance
#'
#' Creates a \code{data.table} of feature importances in a model.
#' Creates a `data.table` of feature importances.
#'
#' @param feature_names character vector of feature names. If the model already
#' contains feature names, those would be used when \code{feature_names=NULL} (default value).
#' Non-null \code{feature_names} could be provided to override those in the model.
#' @param model object of class \code{xgb.Booster}.
#' @param trees (only for the gbtree booster) an integer vector of tree indices that should be included
#' into the importance calculation. If set to \code{NULL}, all trees of the model are parsed.
#' It could be useful, e.g., in multiclass classification to get feature importances
#' for each class separately. IMPORTANT: the tree index in xgboost models
#' is zero-based (e.g., use \code{trees = 0:4} for first 5 trees).
#' @param data deprecated.
#' @param label deprecated.
#' @param target deprecated.
#' @param feature_names Character vector used to overwrite the feature names
#' of the model. The default is `NULL` (use original feature names).
#' @param model Object of class `xgb.Booster`.
#' @param trees An integer vector of tree indices that should be included
#' into the importance calculation (only for the "gbtree" booster).
#' The default (`NULL`) parses all trees.
#' It could be useful, e.g., in multiclass classification to get feature importances
#' for each class separately. *Important*: the tree index in XGBoost models
#' is zero-based (e.g., use `trees = 0:4` for the first five trees).
#' @param data Deprecated.
#' @param label Deprecated.
#' @param target Deprecated.
#'
#' @details
#'
#' This function works for both linear and tree models.
#'
#' For linear models, the importance is the absolute magnitude of linear coefficients.
#' For that reason, in order to obtain a meaningful ranking by importance for a linear model,
#' the features need to be on the same scale (which you also would want to do when using either
#' L1 or L2 regularization).
#' To obtain a meaningful ranking by importance for linear models, the features need to
#' be on the same scale (which is also recommended when using L1 or L2 regularization).
#'
#' @return
#' @return A `data.table` with the following columns:
#'
#' For a tree model, a \code{data.table} with the following columns:
#' \itemize{
#' \item \code{Features} names of the features used in the model;
#' \item \code{Gain} represents fractional contribution of each feature to the model based on
#' the total gain of this feature's splits. Higher percentage means a more important
#' predictive feature.
#' \item \code{Cover} metric of the number of observation related to this feature;
#' \item \code{Frequency} percentage representing the relative number of times
#' a feature have been used in trees.
#' }
#' For a tree model:
#' - `Features`: Names of the features used in the model.
#' - `Gain`: Fractional contribution of each feature to the model based on
#' the total gain of this feature's splits. Higher percentage means higher importance.
#' - `Cover`: Metric of the number of observation related to this feature.
#' - `Frequency`: Percentage of times a feature has been used in trees.
#'
#' A linear model's importance \code{data.table} has the following columns:
#' \itemize{
#' \item \code{Features} names of the features used in the model;
#' \item \code{Weight} the linear coefficient of this feature;
#' \item \code{Class} (only for multiclass models) class label.
#' }
#' For a linear model:
#' - `Features`: Names of the features used in the model.
#' - `Weight`: Linear coefficient of this feature.
#' - `Class`: Class label (only for multiclass models).
#'
#' If \code{feature_names} is not provided and \code{model} doesn't have \code{feature_names},
#' index of the features will be used instead. Because the index is extracted from the model dump
#' If `feature_names` is not provided and `model` doesn't have `feature_names`,
#' the index of the features will be used instead. Because the index is extracted from the model dump
#' (based on C++ code), it starts at 0 (as in C/C++ or Python) instead of 1 (usual in R).
#'
#' @examples
#'
#' # binomial classification using gbtree:
#' data(agaricus.train, package='xgboost')
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
#' eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
#' # binomial classification using "gbtree":
#' data(agaricus.train, package = "xgboost")
#'
#' bst <- xgboost(
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 2,
#' eta = 1,
#' nthread = 2,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#'
#' xgb.importance(model = bst)
#'
#' # binomial classification using gblinear:
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, booster = "gblinear",
#' eta = 0.3, nthread = 1, nrounds = 20, objective = "binary:logistic")
#' # binomial classification using "gblinear":
#' bst <- xgboost(
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' booster = "gblinear",
#' eta = 0.3,
#' nthread = 1,
#' nrounds = 20,objective = "binary:logistic"
#' )
#'
#' xgb.importance(model = bst)
#'
#' # multiclass classification using gbtree:
#' # multiclass classification using "gbtree":
#' nclass <- 3
#' nrounds <- 10
#' mbst <- xgboost(data = as.matrix(iris[, -5]), label = as.numeric(iris$Species) - 1,
#' max_depth = 3, eta = 0.2, nthread = 2, nrounds = nrounds,
#' objective = "multi:softprob", num_class = nclass)
#' mbst <- xgboost(
#' data = as.matrix(iris[, -5]),
#' label = as.numeric(iris$Species) - 1,
#' max_depth = 3,
#' eta = 0.2,
#' nthread = 2,
#' nrounds = nrounds,
#' objective = "multi:softprob",
#' num_class = nclass
#' )
#'
#' # all classes clumped together:
#' xgb.importance(model = mbst)
#' # inspect importances separately for each class:
#' xgb.importance(model = mbst, trees = seq(from=0, by=nclass, length.out=nrounds))
#' xgb.importance(model = mbst, trees = seq(from=1, by=nclass, length.out=nrounds))
#' xgb.importance(model = mbst, trees = seq(from=2, by=nclass, length.out=nrounds))
#'
#' # multiclass classification using gblinear:
#' mbst <- xgboost(data = scale(as.matrix(iris[, -5])), label = as.numeric(iris$Species) - 1,
#' booster = "gblinear", eta = 0.2, nthread = 1, nrounds = 15,
#' objective = "multi:softprob", num_class = nclass)
#' # inspect importances separately for each class:
#' xgb.importance(
#' model = mbst, trees = seq(from = 0, by = nclass, length.out = nrounds)
#' )
#' xgb.importance(
#' model = mbst, trees = seq(from = 1, by = nclass, length.out = nrounds)
#' )
#' xgb.importance(
#' model = mbst, trees = seq(from = 2, by = nclass, length.out = nrounds)
#' )
#'
#' # multiclass classification using "gblinear":
#' mbst <- xgboost(
#' data = scale(as.matrix(iris[, -5])),
#' label = as.numeric(iris$Species) - 1,
#' booster = "gblinear",
#' eta = 0.2,
#' nthread = 1,
#' nrounds = 15,
#' objective = "multi:softprob",
#' num_class = nclass
#' )
#'
#' xgb.importance(model = mbst)
#'
#' @export
@ -87,21 +119,21 @@ xgb.importance <- function(feature_names = NULL, model = NULL, trees = NULL,
if (!(is.null(data) && is.null(label) && is.null(target)))
warning("xgb.importance: parameters 'data', 'label' and 'target' are deprecated")
if (!inherits(model, "xgb.Booster"))
stop("model: must be an object of class xgb.Booster")
if (is.null(feature_names) && !is.null(model$feature_names))
feature_names <- model$feature_names
if (is.null(feature_names)) {
model_feature_names <- xgb.feature_names(model)
if (NROW(model_feature_names)) {
feature_names <- model_feature_names
}
}
if (!(is.null(feature_names) || is.character(feature_names)))
stop("feature_names: Has to be a character vector")
model <- xgb.Booster.complete(model)
config <- jsonlite::fromJSON(xgb.config(model))
if (config$learner$gradient_booster$name == "gblinear") {
handle <- xgb.get.handle(model)
if (xgb.booster_type(model) == "gblinear") {
args <- list(importance_type = "weight", feature_names = feature_names)
results <- .Call(
XGBoosterFeatureScore_R, model$handle, jsonlite::toJSON(args, auto_unbox = TRUE, null = "null")
XGBoosterFeatureScore_R, handle, jsonlite::toJSON(args, auto_unbox = TRUE, null = "null")
)
names(results) <- c("features", "shape", "weight")
if (length(results$shape) == 2) {
@ -122,7 +154,7 @@ xgb.importance <- function(feature_names = NULL, model = NULL, trees = NULL,
for (importance_type in c("weight", "total_gain", "total_cover")) {
args <- list(importance_type = importance_type, feature_names = feature_names, tree_idx = trees)
results <- .Call(
XGBoosterFeatureScore_R, model$handle, jsonlite::toJSON(args, auto_unbox = TRUE, null = "null")
XGBoosterFeatureScore_R, handle, jsonlite::toJSON(args, auto_unbox = TRUE, null = "null")
)
names(results) <- c("features", "shape", importance_type)
concatenated[

27
R-package/R/xgb.load.R
View File

@ -17,7 +17,7 @@
#' An object of \code{xgb.Booster} class.
#'
#' @seealso
#' \code{\link{xgb.save}}, \code{\link{xgb.Booster.complete}}.
#' \code{\link{xgb.save}}
#'
#' @examples
#' data(agaricus.train, package='xgboost')
@ -29,40 +29,37 @@
#'
#' train <- agaricus.train
#' test <- agaricus.test
#' bst <- xgboost(
#' data = train$data, label = train$label, max_depth = 2, eta = 1,
#' bst <- xgb.train(
#' data = xgb.DMatrix(train$data, label = train$label),
#' max_depth = 2,
#' eta = 1,
#' nthread = nthread,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#'
#' xgb.save(bst, 'xgb.model')
#' bst <- xgb.load('xgb.model')
#' if (file.exists('xgb.model')) file.remove('xgb.model')
#' fname <- file.path(tempdir(), "xgb.ubj")
#' xgb.save(bst, fname)
#' bst <- xgb.load(fname)
#' @export
xgb.load <- function(modelfile) {
if (is.null(modelfile))
stop("xgb.load: modelfile cannot be NULL")
handle <- xgb.Booster.handle(
bst <- xgb.Booster(
params = list(),
cachelist = list(),
modelfile = modelfile,
handle = NULL
modelfile = modelfile
)
bst <- bst$bst
# re-use modelfile if it is raw so we do not need to serialize
if (typeof(modelfile) == "raw") {
warning(
paste(
"The support for loading raw booster with `xgb.load` will be ",
"discontinued in upcoming release. Use `xgb.load.raw` or",
" `xgb.unserialize` instead. "
"discontinued in upcoming release. Use `xgb.load.raw` instead. "
)
)
bst <- xgb.handleToBooster(handle = handle, raw = modelfile)
} else {
bst <- xgb.handleToBooster(handle = handle, raw = NULL)
}
bst <- xgb.Booster.complete(bst, saveraw = TRUE)
return(bst)
}

19
R-package/R/xgb.load.raw.R
View File

@ -3,21 +3,10 @@
#' User can generate raw memory buffer by calling xgb.save.raw
#'
#' @param buffer the buffer returned by xgb.save.raw
#' @param as_booster Return the loaded model as xgb.Booster instead of xgb.Booster.handle.
#'
#' @export
xgb.load.raw <- function(buffer, as_booster = FALSE) {
xgb.load.raw <- function(buffer) {
cachelist <- list()
handle <- .Call(XGBoosterCreate_R, cachelist)
.Call(XGBoosterLoadModelFromRaw_R, handle, buffer)
class(handle) <- "xgb.Booster.handle"
if (as_booster) {
booster <- list(handle = handle, raw = NULL)
class(booster) <- "xgb.Booster"
booster <- xgb.Booster.complete(booster, saveraw = TRUE)
return(booster)
} else {
return(handle)
}
bst <- .Call(XGBoosterCreate_R, cachelist)
.Call(XGBoosterLoadModelFromRaw_R, xgb.get.handle(bst), buffer)
return(bst)
}

148
R-package/R/xgb.model.dt.tree.R
View File

@ -1,67 +1,72 @@
#' Parse a boosted tree model text dump
#' Parse model text dump
#'
#' Parse a boosted tree model text dump into a \code{data.table} structure.
#' Parse a boosted tree model text dump into a `data.table` structure.
#'
#' @param feature_names character vector of feature names. If the model already
#' contains feature names, those would be used when \code{feature_names=NULL} (default value).
#' Non-null \code{feature_names} could be provided to override those in the model.
#' @param model object of class \code{xgb.Booster}
#' @param text \code{character} vector previously generated by the \code{xgb.dump}
#' function (where parameter \code{with_stats = TRUE} should have been set).
#' \code{text} takes precedence over \code{model}.
#' @param trees an integer vector of tree indices that should be parsed.
#' If set to \code{NULL}, all trees of the model are parsed.
#' It could be useful, e.g., in multiclass classification to get only
#' the trees of one certain class. IMPORTANT: the tree index in xgboost models
#' is zero-based (e.g., use \code{trees = 0:4} for first 5 trees).
#' @param use_int_id a logical flag indicating whether nodes in columns "Yes", "No", "Missing" should be
#' represented as integers (when FALSE) or as "Tree-Node" character strings (when FALSE).
#' @param ... currently not used.
#' @param feature_names Character vector of feature names. If the model already
#' contains feature names, those will be used when \code{feature_names=NULL} (default value).
#'
#' Note that, if the model already contains feature names, it's \bold{not} possible to override them here.
#' @param model Object of class `xgb.Booster`.
#' @param text Character vector previously generated by the function [xgb.dump()]
#' (called with parameter `with_stats = TRUE`). `text` takes precedence over `model`.
#' @param trees An integer vector of tree indices that should be used.
#' The default (`NULL`) uses all trees.
#' Useful, e.g., in multiclass classification to get only
#' the trees of one class. *Important*: the tree index in XGBoost models
#' is zero-based (e.g., use `trees = 0:4` for the first five trees).
#' @param use_int_id A logical flag indicating whether nodes in columns "Yes", "No", and
#' "Missing" should be represented as integers (when `TRUE`) or as "Tree-Node"
#' character strings (when `FALSE`, default).
#' @param ... Currently not used.
#'
#' @return
#' A \code{data.table} with detailed information about model trees' nodes.
#' A `data.table` with detailed information about tree nodes. It has the following columns:
#' - `Tree`: integer ID of a tree in a model (zero-based index).
#' - `Node`: integer ID of a node in a tree (zero-based index).
#' - `ID`: character identifier of a node in a model (only when `use_int_id = FALSE`).
#' - `Feature`: for a branch node, a feature ID or name (when available);
#' for a leaf node, it simply labels it as `"Leaf"`.
#' - `Split`: location of the split for a branch node (split condition is always "less than").
#' - `Yes`: ID of the next node when the split condition is met.
#' - `No`: ID of the next node when the split condition is not met.
#' - `Missing`: ID of the next node when the branch value is missing.
#' - `Gain`: either the split gain (change in loss) or the leaf value.
#' - `Cover`: metric related to the number of observations either seen by a split
#' or collected by a leaf during training.
#'
#' The columns of the \code{data.table} are:
#'
#' \itemize{
#' \item \code{Tree}: integer ID of a tree in a model (zero-based index)
#' \item \code{Node}: integer ID of a node in a tree (zero-based index)
#' \item \code{ID}: character identifier of a node in a model (only when \code{use_int_id=FALSE})
#' \item \code{Feature}: for a branch node, it's a feature id or name (when available);
#' for a leaf note, it simply labels it as \code{'Leaf'}
#' \item \code{Split}: location of the split for a branch node (split condition is always "less than")
#' \item \code{Yes}: ID of the next node when the split condition is met
#' \item \code{No}: ID of the next node when the split condition is not met
#' \item \code{Missing}: ID of the next node when branch value is missing
#' \item \code{Quality}: either the split gain (change in loss) or the leaf value
#' \item \code{Cover}: metric related to the number of observation either seen by a split
#' or collected by a leaf during training.
#' }
#'
#' When \code{use_int_id=FALSE}, columns "Yes", "No", and "Missing" point to model-wide node identifiers
#' in the "ID" column. When \code{use_int_id=TRUE}, those columns point to node identifiers from
#' When `use_int_id = FALSE`, columns "Yes", "No", and "Missing" point to model-wide node identifiers
#' in the "ID" column. When `use_int_id = TRUE`, those columns point to node identifiers from
#' the corresponding trees in the "Node" column.
#'
#' @examples
#' # Basic use:
#'
#' data(agaricus.train, package='xgboost')
#' data(agaricus.train, package = "xgboost")
#' ## Keep the number of threads to 1 for examples
#' nthread <- 1
#' data.table::setDTthreads(nthread)
#'
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
#' eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
#'
#' (dt <- xgb.model.dt.tree(colnames(agaricus.train$data), bst))
#' bst <- xgboost(
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 2,
#' eta = 1,
#' nthread = nthread,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#'
#' # This bst model already has feature_names stored with it, so those would be used when
#' # feature_names is not set:
#' (dt <- xgb.model.dt.tree(model = bst))
#'
#' # How to match feature names of splits that are following a current 'Yes' branch:
#'
#' merge(dt, dt[, .(ID, Y.Feature=Feature)], by.x='Yes', by.y='ID', all.x=TRUE)[order(Tree,Node)]
#' merge(
#' dt,
#' dt[, .(ID, Y.Feature = Feature)], by.x = "Yes", by.y = "ID", all.x = TRUE
#' )[
#' order(Tree, Node)
#' ]
#'
#' @export
xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
@ -74,8 +79,15 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
" (or NULL if 'model' was provided).")
}
if (is.null(feature_names) && !is.null(model) && !is.null(model$feature_names))
feature_names <- model$feature_names
model_feature_names <- NULL
if (inherits(model, "xgb.Booster")) {
model_feature_names <- xgb.feature_names(model)
if (NROW(model_feature_names) && !is.null(feature_names)) {
stop("'model' contains feature names. Cannot override them.")
}
}
if (is.null(feature_names) && !is.null(model) && !is.null(model_feature_names))
feature_names <- model_feature_names
if (!(is.null(feature_names) || is.character(feature_names))) {
stop("feature_names: must be a character vector")
@ -85,8 +97,10 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
stop("trees: must be a vector of integers.")
}
from_text <- TRUE
if (is.null(text)) {
text <- xgb.dump(model = model, with_stats = TRUE)
from_text <- FALSE
}
if (length(text) < 2 || !any(grepl('leaf=(\\d+)', text))) {
@ -115,9 +129,29 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
td[, isLeaf := grepl("leaf", t, fixed = TRUE)]
# parse branch lines
branch_rx <- paste0("f(\\d+)<(", anynumber_regex, ")\\] yes=(\\d+),no=(\\d+),missing=(\\d+),",
"gain=(", anynumber_regex, "),cover=(", anynumber_regex, ")")
branch_cols <- c("Feature", "Split", "Yes", "No", "Missing", "Quality", "Cover")
branch_rx_nonames <- paste0("f(\\d+)<(", anynumber_regex, ")\\] yes=(\\d+),no=(\\d+),missing=(\\d+),",
"gain=(", anynumber_regex, "),cover=(", anynumber_regex, ")")
branch_rx_w_names <- paste0("\\d+:\\[(.+)<(", anynumber_regex, ")\\] yes=(\\d+),no=(\\d+),missing=(\\d+),",
"gain=(", anynumber_regex, "),cover=(", anynumber_regex, ")")
text_has_feature_names <- FALSE
if (NROW(model_feature_names)) {
branch_rx <- branch_rx_w_names
text_has_feature_names <- TRUE
} else {
# Note: when passing a text dump, it might or might not have feature names,
# but that aspect is unknown from just the text attributes
branch_rx <- branch_rx_nonames
if (from_text) {
if (sum(grepl(branch_rx_w_names, text)) > sum(grepl(branch_rx_nonames, text))) {
branch_rx <- branch_rx_w_names
text_has_feature_names <- TRUE
}
}
}
if (text_has_feature_names && is.null(model) && !is.null(feature_names)) {
stop("'text' contains feature names. Cannot override them.")
}
branch_cols <- c("Feature", "Split", "Yes", "No", "Missing", "Gain", "Cover")
td[
isLeaf == FALSE,
(branch_cols) := {
@ -127,7 +161,7 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
xtr[, 3:5] <- add.tree.id(xtr[, 3:5], Tree)
if (length(xtr) == 0) {
as.data.table(
list(Feature = "NA", Split = "NA", Yes = "NA", No = "NA", Missing = "NA", Quality = "NA", Cover = "NA")
list(Feature = "NA", Split = "NA", Yes = "NA", No = "NA", Missing = "NA", Gain = "NA", Cover = "NA")
)
} else {
as.data.table(xtr)
@ -139,15 +173,17 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
is_stump <- function() {
return(length(td$Feature) == 1 && is.na(td$Feature))
}
if (!is.null(feature_names) && !is_stump()) {
if (length(feature_names) <= max(as.numeric(td$Feature), na.rm = TRUE))
stop("feature_names has less elements than there are features used in the model")
td[isLeaf == FALSE, Feature := feature_names[as.numeric(Feature) + 1]]
if (!text_has_feature_names) {
if (!is.null(feature_names) && !is_stump()) {
if (length(feature_names) <= max(as.numeric(td$Feature), na.rm = TRUE))
stop("feature_names has less elements than there are features used in the model")
td[isLeaf == FALSE, Feature := feature_names[as.numeric(Feature) + 1]]
}
}
# parse leaf lines
leaf_rx <- paste0("leaf=(", anynumber_regex, "),cover=(", anynumber_regex, ")")
leaf_cols <- c("Feature", "Quality", "Cover")
leaf_cols <- c("Feature", "Gain", "Cover")
td[
isLeaf == TRUE,
(leaf_cols) := {
@ -162,7 +198,7 @@ xgb.model.dt.tree <- function(feature_names = NULL, model = NULL, text = NULL,
]
# convert some columns to numeric
numeric_cols <- c("Split", "Quality", "Cover")
numeric_cols <- c("Split", "Gain", "Cover")
td[, (numeric_cols) := lapply(.SD, as.numeric), .SDcols = numeric_cols]
if (use_int_id) {
int_cols <- c("Yes", "No", "Missing")

87
R-package/R/xgb.plot.deepness.R
View File

@ -1,65 +1,74 @@
#' Plot model trees deepness
#' Plot model tree depth
#'
#' Visualizes distributions related to depth of tree leafs.
#' \code{xgb.plot.deepness} uses base R graphics, while \code{xgb.ggplot.deepness} uses the ggplot backend.
#' Visualizes distributions related to the depth of tree leaves.
#' - `xgb.plot.deepness()` uses base R graphics, while
#' - `xgb.ggplot.deepness()` uses "ggplot2".
#'
#' @param model either an \code{xgb.Booster} model generated by the \code{xgb.train} function
#' or a data.table result of the \code{xgb.model.dt.tree} function.
#' @param plot (base R barplot) whether a barplot should be produced.
#' If FALSE, only a data.table is returned.
#' @param which which distribution to plot (see details).
#' @param ... other parameters passed to \code{barplot} or \code{plot}.
#' @param model Either an `xgb.Booster` model, or the "data.table" returned by [xgb.model.dt.tree()].
#' @param which Which distribution to plot (see details).
#' @param plot Should the plot be shown? Default is `TRUE`.
#' @param ... Other parameters passed to [graphics::barplot()] or [graphics::plot()].
#'
#' @details
#'
#' When \code{which="2x1"}, two distributions with respect to the leaf depth
#' When `which = "2x1"`, two distributions with respect to the leaf depth
#' are plotted on top of each other:
#' \itemize{
#' \item the distribution of the number of leafs in a tree model at a certain depth;
#' \item the distribution of average weighted number of observations ("cover")
#' ending up in leafs at certain depth.
#' }
#' Those could be helpful in determining sensible ranges of the \code{max_depth}
#' and \code{min_child_weight} parameters.
#' 1. The distribution of the number of leaves in a tree model at a certain depth.
#' 2. The distribution of the average weighted number of observations ("cover")
#' ending up in leaves at a certain depth.
#'
#' When \code{which="max.depth"} or \code{which="med.depth"}, plots of either maximum or median depth
#' per tree with respect to tree number are created. And \code{which="med.weight"} allows to see how
#' Those could be helpful in determining sensible ranges of the `max_depth`
#' and `min_child_weight` parameters.
#'
#' When `which = "max.depth"` or `which = "med.depth"`, plots of either maximum or
#' median depth per tree with respect to the tree number are created.
#'
#' Finally, `which = "med.weight"` allows to see how
#' a tree's median absolute leaf weight changes through the iterations.
#'
#' This function was inspired by the blog post
#' \url{https://github.com/aysent/random-forest-leaf-visualization}.
#' These functions have been inspired by the blog post
#' <https://github.com/aysent/random-forest-leaf-visualization>.
#'
#' @return
#' The return value of the two functions is as follows:
#' - `xgb.plot.deepness()`: A "data.table" (invisibly).
#' Each row corresponds to a terminal leaf in the model. It contains its information
#' about depth, cover, and weight (used in calculating predictions).
#' If `plot = TRUE`, also a plot is shown.
#' - `xgb.ggplot.deepness()`: When `which = "2x1"`, a list of two "ggplot" objects,
#' and a single "ggplot" object otherwise.
#'
#' Other than producing plots (when \code{plot=TRUE}), the \code{xgb.plot.deepness} function
#' silently returns a processed data.table where each row corresponds to a terminal leaf in a tree model,
#' and contains information about leaf's depth, cover, and weight (which is used in calculating predictions).
#'
#' The \code{xgb.ggplot.deepness} silently returns either a list of two ggplot graphs when \code{which="2x1"}
#' or a single ggplot graph for the other \code{which} options.
#'
#' @seealso
#'
#' \code{\link{xgb.train}}, \code{\link{xgb.model.dt.tree}}.
#' @seealso [xgb.train()] and [xgb.model.dt.tree()].
#'
#' @examples
#'
#' data(agaricus.train, package='xgboost')
#' data(agaricus.train, package = "xgboost")
#' ## Keep the number of threads to 2 for examples
#' nthread <- 2
#' data.table::setDTthreads(nthread)
#'
#' ## Change max_depth to a higher number to get a more significant result
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 6,
#' eta = 0.1, nthread = nthread, nrounds = 50, objective = "binary:logistic",
#' subsample = 0.5, min_child_weight = 2)
#' bst <- xgboost(
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 6,
#' nthread = nthread,
#' nrounds = 50,
#' objective = "binary:logistic",
#' subsample = 0.5,
#' min_child_weight = 2
#' )
#'
#' xgb.plot.deepness(bst)
#' xgb.ggplot.deepness(bst)
#'
#' xgb.plot.deepness(bst, which='max.depth', pch=16, col=rgb(0,0,1,0.3), cex=2)
#' xgb.plot.deepness(
#' bst, which = "max.depth", pch = 16, col = rgb(0, 0, 1, 0.3), cex = 2
#' )
#'
#' xgb.plot.deepness(bst, which='med.weight', pch=16, col=rgb(0,0,1,0.3), cex=2)
#' xgb.plot.deepness(
#' bst, which = "med.weight", pch = 16, col = rgb(0, 0, 1, 0.3), cex = 2
#' )
#'
#' @rdname xgb.plot.deepness
#' @export
@ -83,7 +92,7 @@ xgb.plot.deepness <- function(model = NULL, which = c("2x1", "max.depth", "med.d
stop("Model tree columns are not as expected!\n",
" Note that this function works only for tree models.")
dt_depths <- merge(get.leaf.depth(dt_tree), dt_tree[, .(ID, Cover, Weight = Quality)], by = "ID")
dt_depths <- merge(get.leaf.depth(dt_tree), dt_tree[, .(ID, Cover, Weight = Gain)], by = "ID")
setkeyv(dt_depths, c("Tree", "ID"))
# count by depth levels, and also calculate average cover at a depth
dt_summaries <- dt_depths[, .(.N, Cover = mean(Cover)), Depth]
@ -148,6 +157,6 @@ get.leaf.depth <- function(dt_tree) {
# They are mainly column names inferred by Data.table...
globalVariables(
c(
".N", "N", "Depth", "Quality", "Cover", "Tree", "ID", "Yes", "No", "Feature", "Leaf", "Weight"
".N", "N", "Depth", "Gain", "Cover", "Tree", "ID", "Yes", "No", "Feature", "Leaf", "Weight"
)
)

81
R-package/R/xgb.plot.importance.R
View File

@ -1,64 +1,75 @@
#' Plot feature importance as a bar graph
#' Plot feature importance
#'
#' Represents previously calculated feature importance as a bar graph.
#' \code{xgb.plot.importance} uses base R graphics, while \code{xgb.ggplot.importance} uses the ggplot backend.
#' - `xgb.plot.importance()` uses base R graphics, while
#' - `xgb.ggplot.importance()` uses "ggplot".
#'
#' @param importance_matrix a \code{data.table} returned by \code{\link{xgb.importance}}.
#' @param top_n maximal number of top features to include into the plot.
#' @param measure the name of importance measure to plot.
#' When \code{NULL}, 'Gain' would be used for trees and 'Weight' would be used for gblinear.
#' @param rel_to_first whether importance values should be represented as relative to the highest ranked feature.
#' See Details.
#' @param left_margin (base R barplot) allows to adjust the left margin size to fit feature names.
#' When it is NULL, the existing \code{par('mar')} is used.
#' @param cex (base R barplot) passed as \code{cex.names} parameter to \code{barplot}.
#' @param plot (base R barplot) whether a barplot should be produced.
#' If FALSE, only a data.table is returned.
#' @param n_clusters (ggplot only) a \code{numeric} vector containing the min and the max range
#' @param importance_matrix A `data.table` as returned by [xgb.importance()].
#' @param top_n Maximal number of top features to include into the plot.
#' @param measure The name of importance measure to plot.
#' When `NULL`, 'Gain' would be used for trees and 'Weight' would be used for gblinear.
#' @param rel_to_first Whether importance values should be represented as relative to
#' the highest ranked feature, see Details.
#' @param left_margin Adjust the left margin size to fit feature names.
#' When `NULL`, the existing `par("mar")` is used.
#' @param cex Passed as `cex.names` parameter to [graphics::barplot()].
#' @param plot Should the barplot be shown? Default is `TRUE`.
#' @param n_clusters A numeric vector containing the min and the max range
#' of the possible number of clusters of bars.
#' @param ... other parameters passed to \code{barplot} (except horiz, border, cex.names, names.arg, and las).
#' @param ... Other parameters passed to [graphics::barplot()]
#' (except `horiz`, `border`, `cex.names`, `names.arg`, and `las`).
#' Only used in `xgb.plot.importance()`.
#'
#' @details
#' The graph represents each feature as a horizontal bar of length proportional to the importance of a feature.
#' Features are shown ranked in a decreasing importance order.
#' It works for importances from both \code{gblinear} and \code{gbtree} models.
#' Features are sorted by decreasing importance.
#' It works for both "gblinear" and "gbtree" models.
#'
#' When \code{rel_to_first = FALSE}, the values would be plotted as they were in \code{importance_matrix}.
#' For gbtree model, that would mean being normalized to the total of 1
#' When `rel_to_first = FALSE`, the values would be plotted as in `importance_matrix`.
#' For a "gbtree" model, that would mean being normalized to the total of 1
#' ("what is feature's importance contribution relative to the whole model?").
#' For linear models, \code{rel_to_first = FALSE} would show actual values of the coefficients.
#' Setting \code{rel_to_first = TRUE} allows to see the picture from the perspective of
#' For linear models, `rel_to_first = FALSE` would show actual values of the coefficients.
#' Setting `rel_to_first = TRUE` allows to see the picture from the perspective of
#' "what is feature's importance contribution relative to the most important feature?"
#'
#' The ggplot-backend method also performs 1-D clustering of the importance values,
#' with bar colors corresponding to different clusters that have somewhat similar importance values.
#' The "ggplot" backend performs 1-D clustering of the importance values,
#' with bar colors corresponding to different clusters having similar importance values.
#'
#' @return
#' The \code{xgb.plot.importance} function creates a \code{barplot} (when \code{plot=TRUE})
#' and silently returns a processed data.table with \code{n_top} features sorted by importance.
#' The return value depends on the function:
#' - `xgb.plot.importance()`: Invisibly, a "data.table" with `n_top` features sorted
#' by importance. If `plot = TRUE`, the values are also plotted as barplot.
#' - `xgb.ggplot.importance()`: A customizable "ggplot" object.
#' E.g., to change the title, set `+ ggtitle("A GRAPH NAME")`.
#'
#' The \code{xgb.ggplot.importance} function returns a ggplot graph which could be customized afterwards.
#' E.g., to change the title of the graph, add \code{+ ggtitle("A GRAPH NAME")} to the result.
#'
#' @seealso
#' \code{\link[graphics]{barplot}}.
#' @seealso [graphics::barplot()]
#'
#' @examples
#' data(agaricus.train)
#'
#' ## Keep the number of threads to 2 for examples
#' nthread <- 2
#' data.table::setDTthreads(nthread)
#'
#' bst <- xgboost(
#' data = agaricus.train$data, label = agaricus.train$label, max_depth = 3,
#' eta = 1, nthread = nthread, nrounds = 2, objective = "binary:logistic"
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 3,
#' eta = 1,
#' nthread = nthread,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#'
#' importance_matrix <- xgb.importance(colnames(agaricus.train$data), model = bst)
#' xgb.plot.importance(
#' importance_matrix, rel_to_first = TRUE, xlab = "Relative importance"
#' )
#'
#' xgb.plot.importance(importance_matrix, rel_to_first = TRUE, xlab = "Relative importance")
#'
#' (gg <- xgb.ggplot.importance(importance_matrix, measure = "Frequency", rel_to_first = TRUE))
#' gg <- xgb.ggplot.importance(
#' importance_matrix, measure = "Frequency", rel_to_first = TRUE
#' )
#' gg
#' gg + ggplot2::ylab("Frequency")
#'
#' @rdname xgb.plot.importance

55
R-package/R/xgb.plot.multi.trees.R
View File

@ -1,14 +1,10 @@
#' Project all trees on one tree and plot it
#' Project all trees on one tree
#'
#' Visualization of the ensemble of trees as a single collective unit.
#'
#' @param model produced by the \code{xgb.train} function.
#' @param feature_names names of each feature as a \code{character} vector.
#' @param features_keep number of features to keep in each position of the multi trees.
#' @param plot_width width in pixels of the graph to produce
#' @param plot_height height in pixels of the graph to produce
#' @param render a logical flag for whether the graph should be rendered (see Value).
#' @param ... currently not used
#' @inheritParams xgb.plot.tree
#' @param features_keep Number of features to keep in each position of the multi trees,
#' by default 5.
#'
#' @details
#'
@ -24,33 +20,31 @@
#' Moreover, the trees tend to reuse the same features.
#'
#' The function projects each tree onto one, and keeps for each position the
#' \code{features_keep} first features (based on the Gain per feature measure).
#' `features_keep` first features (based on the Gain per feature measure).
#'
#' This function is inspired by this blog post:
#' \url{https://wellecks.wordpress.com/2015/02/21/peering-into-the-black-box-visualizing-lambdamart/}
#' <https://wellecks.wordpress.com/2015/02/21/peering-into-the-black-box-visualizing-lambdamart/>
#'
#' @return
#'
#' When \code{render = TRUE}:
#' returns a rendered graph object which is an \code{htmlwidget} of class \code{grViz}.
#' Similar to ggplot objects, it needs to be printed to see it when not running from command line.
#'
#' When \code{render = FALSE}:
#' silently returns a graph object which is of DiagrammeR's class \code{dgr_graph}.
#' This could be useful if one wants to modify some of the graph attributes
#' before rendering the graph with \code{\link[DiagrammeR]{render_graph}}.
#' @inherit xgb.plot.tree return
#'
#' @examples
#'
#' data(agaricus.train, package='xgboost')
#' data(agaricus.train, package = "xgboost")
#'
#' ## Keep the number of threads to 2 for examples
#' nthread <- 2
#' data.table::setDTthreads(nthread)
#'
#' bst <- xgboost(
#' data = agaricus.train$data, label = agaricus.train$label, max_depth = 15,
#' eta = 1, nthread = nthread, nrounds = 30, objective = "binary:logistic",
#' min_child_weight = 50, verbose = 0
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 15,
#' eta = 1,
#' nthread = nthread,
#' nrounds = 30,
#' objective = "binary:logistic",
#' min_child_weight = 50,
#' verbose = 0
#' )
#'
#' p <- xgb.plot.multi.trees(model = bst, features_keep = 3)
@ -58,10 +52,13 @@
#'
#' \dontrun{
#' # Below is an example of how to save this plot to a file.
#' # Note that for `export_graph` to work, the DiagrammeRsvg and rsvg packages must also be installed.
#' # Note that for export_graph() to work, the {DiagrammeRsvg} and {rsvg} packages
#' # must also be installed.
#'
#' library(DiagrammeR)
#' gr <- xgb.plot.multi.trees(model=bst, features_keep = 3, render=FALSE)
#' export_graph(gr, 'tree.pdf', width=1500, height=600)
#'
#' gr <- xgb.plot.multi.trees(model = bst, features_keep = 3, render = FALSE)
#' export_graph(gr, "tree.pdf", width = 1500, height = 600)
#' }
#'
#' @export
@ -98,13 +95,13 @@ xgb.plot.multi.trees <- function(model, feature_names = NULL, features_keep = 5,
data.table::set(tree.matrix, j = nm, value = sub("^\\d+-", "", tree.matrix[[nm]]))
nodes.dt <- tree.matrix[
, .(Quality = sum(Quality))
, .(Gain = sum(Gain))
, by = .(abs.node.position, Feature)
][, .(Text = paste0(
paste0(
Feature[seq_len(min(length(Feature), features_keep))],
" (",
format(Quality[seq_len(min(length(Quality), features_keep))], digits = 5),
format(Gain[seq_len(min(length(Gain), features_keep))], digits = 5),
")"
),
collapse = "\n"

244
R-package/R/xgb.plot.shap.R
View File

@ -1,110 +1,165 @@
#' SHAP contribution dependency plots
#' SHAP dependence plots
#'
#' Visualizing the SHAP feature contribution to prediction dependencies on feature value.
#' Visualizes SHAP values against feature values to gain an impression of feature effects.
#'
#' @param data data as a \code{matrix} or \code{dgCMatrix}.
#' @param shap_contrib a matrix of SHAP contributions that was computed earlier for the above
#' \code{data}. When it is NULL, it is computed internally using \code{model} and \code{data}.
#' @param features a vector of either column indices or of feature names to plot. When it is NULL,
#' feature importance is calculated, and \code{top_n} high ranked features are taken.
#' @param top_n when \code{features} is NULL, top_n [1, 100] most important features in a model are taken.
#' @param model an \code{xgb.Booster} model. It has to be provided when either \code{shap_contrib}
#' or \code{features} is missing.
#' @param trees passed to \code{\link{xgb.importance}} when \code{features = NULL}.
#' @param target_class is only relevant for multiclass models. When it is set to a 0-based class index,
#' only SHAP contributions for that specific class are used.
#' If it is not set, SHAP importances are averaged over all classes.
#' @param approxcontrib passed to \code{\link{predict.xgb.Booster}} when \code{shap_contrib = NULL}.
#' @param subsample a random fraction of data points to use for plotting. When it is NULL,
#' it is set so that up to 100K data points are used.
#' @param n_col a number of columns in a grid of plots.
#' @param col color of the scatterplot markers.
#' @param pch scatterplot marker.
#' @param discrete_n_uniq a maximal number of unique values in a feature to consider it as discrete.
#' @param discrete_jitter an \code{amount} parameter of jitter added to discrete features' positions.
#' @param ylab a y-axis label in 1D plots.
#' @param plot_NA whether the contributions of cases with missing values should also be plotted.
#' @param col_NA a color of marker for missing value contributions.
#' @param pch_NA a marker type for NA values.
#' @param pos_NA a relative position of the x-location where NA values are shown:
#' \code{min(x) + (max(x) - min(x)) * pos_NA}.
#' @param plot_loess whether to plot loess-smoothed curves. The smoothing is only done for features with
#' more than 5 distinct values.
#' @param col_loess a color to use for the loess curves.
#' @param span_loess the \code{span} parameter in \code{\link[stats]{loess}}'s call.
#' @param which whether to do univariate or bivariate plotting. NOTE: only 1D is implemented so far.
#' @param plot whether a plot should be drawn. If FALSE, only a list of matrices is returned.
#' @param ... other parameters passed to \code{plot}.
#' @param data The data to explain as a `matrix` or `dgCMatrix`.
#' @param shap_contrib Matrix of SHAP contributions of `data`.
#' The default (`NULL`) computes it from `model` and `data`.
#' @param features Vector of column indices or feature names to plot.
#' When `NULL` (default), the `top_n` most important features are selected
#' by [xgb.importance()].
#' @param top_n How many of the most important features (<= 100) should be selected?
#' By default 1 for SHAP dependence and 10 for SHAP summary).
#' Only used when `features = NULL`.
#' @param model An `xgb.Booster` model. Only required when `shap_contrib = NULL` or
#' `features = NULL`.
#' @param trees Passed to [xgb.importance()] when `features = NULL`.
#' @param target_class Only relevant for multiclass models. The default (`NULL`)
#' averages the SHAP values over all classes. Pass a (0-based) class index
#' to show only SHAP values of that class.
#' @param approxcontrib Passed to `predict()` when `shap_contrib = NULL`.
#' @param subsample Fraction of data points randomly picked for plotting.
#' The default (`NULL`) will use up to 100k data points.
#' @param n_col Number of columns in a grid of plots.
#' @param col Color of the scatterplot markers.
#' @param pch Scatterplot marker.
#' @param discrete_n_uniq Maximal number of unique feature values to consider the
#' feature as discrete.
#' @param discrete_jitter Jitter amount added to the values of discrete features.
#' @param ylab The y-axis label in 1D plots.
#' @param plot_NA Should contributions of cases with missing values be plotted?
#' Default is `TRUE`.
#' @param col_NA Color of marker for missing value contributions.
#' @param pch_NA Marker type for `NA` values.
#' @param pos_NA Relative position of the x-location where `NA` values are shown:
#' `min(x) + (max(x) - min(x)) * pos_NA`.
#' @param plot_loess Should loess-smoothed curves be plotted? (Default is `TRUE`).
#' The smoothing is only done for features with more than 5 distinct values.
#' @param col_loess Color of loess curves.
#' @param span_loess The `span` parameter of [stats::loess()].
#' @param which Whether to do univariate or bivariate plotting. Currently, only "1d" is implemented.
#' @param plot Should the plot be drawn? (Default is `TRUE`).
#' If `FALSE`, only a list of matrices is returned.
#' @param ... Other parameters passed to [graphics::plot()].
#'
#' @details
#'
#' These scatterplots represent how SHAP feature contributions depend of feature values.
#' The similarity to partial dependency plots is that they also give an idea for how feature values
#' affect predictions. However, in partial dependency plots, we usually see marginal dependencies
#' of model prediction on feature value, while SHAP contribution dependency plots display the estimated
#' contributions of a feature to model prediction for each individual case.
#' The similarity to partial dependence plots is that they also give an idea for how feature values
#' affect predictions. However, in partial dependence plots, we see marginal dependencies
#' of model prediction on feature value, while SHAP dependence plots display the estimated
#' contributions of a feature to the prediction for each individual case.
#'
#' When \code{plot_loess = TRUE} is set, feature values are rounded to 3 significant digits and
#' weighted LOESS is computed and plotted, where weights are the numbers of data points
#' When `plot_loess = TRUE`, feature values are rounded to three significant digits and
#' weighted LOESS is computed and plotted, where the weights are the numbers of data points
#' at each rounded value.
#'
#' Note: SHAP contributions are shown on the scale of model margin. E.g., for a logistic binomial objective,
#' the margin is prediction before a sigmoidal transform into probability-like values.
#' Note: SHAP contributions are on the scale of the model margin.
#' E.g., for a logistic binomial objective, the margin is on log-odds scale.
#' Also, since SHAP stands for "SHapley Additive exPlanation" (model prediction = sum of SHAP
#' contributions for all features + bias), depending on the objective used, transforming SHAP
#' contributions for a feature from the marginal to the prediction space is not necessarily
#' a meaningful thing to do.
#'
#' @return
#'
#' In addition to producing plots (when \code{plot=TRUE}), it silently returns a list of two matrices:
#' \itemize{
#' \item \code{data} the values of selected features;
#' \item \code{shap_contrib} the contributions of selected features.
#' }
#' In addition to producing plots (when `plot = TRUE`), it silently returns a list of two matrices:
#' - `data`: Feature value matrix.
#' - `shap_contrib`: Corresponding SHAP value matrix.
#'
#' @references
#'
#' Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions", NIPS Proceedings 2017, \url{https://arxiv.org/abs/1705.07874}
#'
#' Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles", \url{https://arxiv.org/abs/1706.06060}
#' 1. Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions",
#' NIPS Proceedings 2017, <https://arxiv.org/abs/1705.07874>
#' 2. Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles",
#' <https://arxiv.org/abs/1706.06060>
#'
#' @examples
#'
#' data(agaricus.train, package='xgboost')
#' data(agaricus.test, package='xgboost')
#' data(agaricus.train, package = "xgboost")
#' data(agaricus.test, package = "xgboost")
#'
#' ## Keep the number of threads to 1 for examples
#' nthread <- 1
#' data.table::setDTthreads(nthread)
#' nrounds <- 20
#'
#' bst <- xgboost(agaricus.train$data, agaricus.train$label, nrounds = nrounds,
#' eta = 0.1, max_depth = 3, subsample = .5,
#' method = "hist", objective = "binary:logistic", nthread = nthread, verbose = 0)
#' bst <- xgboost(
#' agaricus.train$data,
#' agaricus.train$label,
#' nrounds = nrounds,
#' eta = 0.1,
#' max_depth = 3,
#' subsample = 0.5,
#' objective = "binary:logistic",
#' nthread = nthread,
#' verbose = 0
#' )
#'
#' xgb.plot.shap(agaricus.test$data, model = bst, features = "odor=none")
#'
#' contr <- predict(bst, agaricus.test$data, predcontrib = TRUE)
#' xgb.plot.shap(agaricus.test$data, contr, model = bst, top_n = 12, n_col = 3)
#' xgb.ggplot.shap.summary(agaricus.test$data, contr, model = bst, top_n = 12) # Summary plot
#'
#' # multiclass example - plots for each class separately:
#' # Summary plot
#' xgb.ggplot.shap.summary(agaricus.test$data, contr, model = bst, top_n = 12)
#'
#' # Multiclass example - plots for each class separately:
#' nclass <- 3
#' x <- as.matrix(iris[, -5])
#' set.seed(123)
#' is.na(x[sample(nrow(x) * 4, 30)]) <- TRUE # introduce some missing values
#' mbst <- xgboost(data = x, label = as.numeric(iris$Species) - 1, nrounds = nrounds,
#' max_depth = 2, eta = 0.3, subsample = .5, nthread = nthread,
#' objective = "multi:softprob", num_class = nclass, verbose = 0)
#' trees0 <- seq(from=0, by=nclass, length.out=nrounds)
#'
#' mbst <- xgboost(
#' data = x,
#' label = as.numeric(iris$Species) - 1,
#' nrounds = nrounds,
#' max_depth = 2,
#' eta = 0.3,
#' subsample = 0.5,
#' nthread = nthread,
#' objective = "multi:softprob",
#' num_class = nclass,
#' verbose = 0
#' )
#' trees0 <- seq(from = 0, by = nclass, length.out = nrounds)
#' col <- rgb(0, 0, 1, 0.5)
#' xgb.plot.shap(x, model = mbst, trees = trees0, target_class = 0, top_n = 4,
#' n_col = 2, col = col, pch = 16, pch_NA = 17)
#' xgb.plot.shap(x, model = mbst, trees = trees0 + 1, target_class = 1, top_n = 4,
#' n_col = 2, col = col, pch = 16, pch_NA = 17)
#' xgb.plot.shap(x, model = mbst, trees = trees0 + 2, target_class = 2, top_n = 4,
#' n_col = 2, col = col, pch = 16, pch_NA = 17)
#' xgb.ggplot.shap.summary(x, model = mbst, target_class = 0, top_n = 4) # Summary plot
#' xgb.plot.shap(
#' x,
#' model = mbst,
#' trees = trees0,
#' target_class = 0,
#' top_n = 4,
#' n_col = 2,
#' col = col,
#' pch = 16,
#' pch_NA = 17
#' )
#'
#' xgb.plot.shap(
#' x,
#' model = mbst,
#' trees = trees0 + 1,
#' target_class = 1,
#' top_n = 4,
#' n_col = 2,
#' col = col,
#' pch = 16,
#' pch_NA = 17
#' )
#'
#' xgb.plot.shap(
#' x,
#' model = mbst,
#' trees = trees0 + 2,
#' target_class = 2,
#' top_n = 4,
#' n_col = 2,
#' col = col,
#' pch = 16,
#' pch_NA = 17
#' )
#'
#' # Summary plot
#' xgb.ggplot.shap.summary(x, model = mbst, target_class = 0, top_n = 4)
#'
#' @rdname xgb.plot.shap
#' @export
@ -187,41 +242,48 @@ xgb.plot.shap <- function(data, shap_contrib = NULL, features = NULL, top_n = 1,
invisible(list(data = data, shap_contrib = shap_contrib))
}
#' SHAP contribution dependency summary plot
#' SHAP summary plot
#'
#' Compare SHAP contributions of different features.
#' Visualizes SHAP contributions of different features.
#'
#' A point plot (each point representing one sample from \code{data}) is
#' A point plot (each point representing one observation from `data`) is
#' produced for each feature, with the points plotted on the SHAP value axis.
#' Each point (observation) is coloured based on its feature value. The plot
#' hence allows us to see which features have a negative / positive contribution
#' Each point (observation) is coloured based on its feature value.
#'
#' The plot allows to see which features have a negative / positive contribution
#' on the model prediction, and whether the contribution is different for larger
#' or smaller values of the feature. We effectively try to replicate the
#' \code{summary_plot} function from https://github.com/shap/shap.
#' or smaller values of the feature. Inspired by the summary plot of
#' <https://github.com/shap/shap>.
#'
#' @inheritParams xgb.plot.shap
#'
#' @return A \code{ggplot2} object.
#' @return A `ggplot2` object.
#' @export
#'
#' @examples # See \code{\link{xgb.plot.shap}}.
#' @seealso \code{\link{xgb.plot.shap}}, \code{\link{xgb.ggplot.shap.summary}},
#' \url{https://github.com/shap/shap}
#' @examples
#' # See examples in xgb.plot.shap()
#'
#' @seealso [xgb.plot.shap()], [xgb.ggplot.shap.summary()],
#' and the Python library <https://github.com/shap/shap>.
xgb.plot.shap.summary <- function(data, shap_contrib = NULL, features = NULL, top_n = 10, model = NULL,
trees = NULL, target_class = NULL, approxcontrib = FALSE, subsample = NULL) {
# Only ggplot implementation is available.
xgb.ggplot.shap.summary(data, shap_contrib, features, top_n, model, trees, target_class, approxcontrib, subsample)
}
#' Prepare data for SHAP plots. To be used in xgb.plot.shap, xgb.plot.shap.summary, etc.
#' Internal utility function.
#' Prepare data for SHAP plots
#'
#' Internal function used in [xgb.plot.shap()], [xgb.plot.shap.summary()], etc.
#'
#' @inheritParams xgb.plot.shap
#' @param max_observations Maximum number of observations to consider.
#' @keywords internal
#' @noRd
#'
#' @return A list containing: 'data', a matrix containing sample observations
#' and their feature values; 'shap_contrib', a matrix containing the SHAP contribution
#' values for these observations.
#' @return
#' A list containing:
#' - `data`: The matrix of feature values.
#' - `shap_contrib`: The matrix with corresponding SHAP values.
xgb.shap.data <- function(data, shap_contrib = NULL, features = NULL, top_n = 1, model = NULL,
trees = NULL, target_class = NULL, approxcontrib = FALSE,
subsample = NULL, max_observations = 100000) {
@ -241,7 +303,11 @@ xgb.shap.data <- function(data, shap_contrib = NULL, features = NULL, top_n = 1,
if (is.character(features) && is.null(colnames(data)))
stop("either provide `data` with column names or provide `features` as column indices")
if (is.null(model$feature_names) && model$nfeatures != ncol(data))
model_feature_names <- NULL
if (is.null(features) && !is.null(model)) {
model_feature_names <- xgb.feature_names(model)
}
if (is.null(model_feature_names) && xgb.num_feature(model) != ncol(data))
stop("if model has no feature_names, columns in `data` must match features in model")
if (!is.null(subsample)) {
@ -270,7 +336,7 @@ xgb.shap.data <- function(data, shap_contrib = NULL, features = NULL, top_n = 1,
}
if (is.null(features)) {
if (!is.null(model$feature_names)) {
if (!is.null(model_feature_names)) {
imp <- xgb.importance(model = model, trees = trees)
} else {
imp <- xgb.importance(model = model, trees = trees, feature_names = colnames(data))

138
R-package/R/xgb.plot.tree.R
View File

@ -1,74 +1,110 @@
#' Plot a boosted tree model
#' Plot boosted trees
#'
#' Read a tree model text dump and plot the model.
#'
#' @param feature_names names of each feature as a \code{character} vector.
#' @param model produced by the \code{xgb.train} function.
#' @param trees an integer vector of tree indices that should be visualized.
#' If set to \code{NULL}, all trees of the model are included.
#' IMPORTANT: the tree index in xgboost model is zero-based
#' (e.g., use \code{trees = 0:2} for the first 3 trees in a model).
#' @param plot_width the width of the diagram in pixels.
#' @param plot_height the height of the diagram in pixels.
#' @param render a logical flag for whether the graph should be rendered (see Value).
#' @param feature_names Character vector used to overwrite the feature names
#' of the model. The default (`NULL`) uses the original feature names.
#' @param model Object of class `xgb.Booster`.
#' @param trees An integer vector of tree indices that should be used.
#' The default (`NULL`) uses all trees.
#' Useful, e.g., in multiclass classification to get only
#' the trees of one class. *Important*: the tree index in XGBoost models
#' is zero-based (e.g., use `trees = 0:2` for the first three trees).
#' @param plot_width,plot_height Width and height of the graph in pixels.
#' The values are passed to [DiagrammeR::render_graph()].
#' @param render Should the graph be rendered or not? The default is `TRUE`.
#' @param show_node_id a logical flag for whether to show node id's in the graph.
#' @param style Style to use for the plot. Options are:\itemize{
#' \item `"xgboost"`: will use the plot style defined in the core XGBoost library,
#' which is shared between different interfaces through the 'dot' format. This
#' style was not available before version 2.1.0 in R. It always plots the trees
#' vertically (from top to bottom).
#' \item `"R"`: will use the style defined from XGBoost's R interface, which predates
#' the introducition of the standardized style from the core library. It might plot
#' the trees horizontally (from left to right).
#' }
#'
#' Note that `style="xgboost"` is only supported when all of the following conditions are met:\itemize{
#' \item Only a single tree is being plotted.
#' \item Node IDs are not added to the graph.
#' \item The graph is being returned as `htmlwidget` (`render=TRUE`).
#' }
#' @param ... currently not used.
#'
#' @details
#'
#' The content of each node is organised that way:
#' When using `style="xgboost"`, the content of each node is visualized as follows:
#' - For non-terminal nodes, it will display the split condition (number or name if
#' available, and the condition that would decide to which node to go next).
#' - Those nodes will be connected to their children by arrows that indicate whether the
#' branch corresponds to the condition being met or not being met.
#' - Terminal (leaf) nodes contain the margin to add when ending there.
#'
#' \itemize{
#' \item Feature name.
#' \item \code{Cover}: The sum of second order gradient of training data classified to the leaf.
#' If it is square loss, this simply corresponds to the number of instances seen by a split
#' or collected by a leaf during training.
#' The deeper in the tree a node is, the lower this metric will be.
#' \item \code{Gain} (for split nodes): the information gain metric of a split
#' When using `style="R"`, the content of each node is visualized like this:
#' - *Feature name*.
#' - *Cover:* The sum of second order gradients of training data.
#' For the squared loss, this simply corresponds to the number of instances in the node.
#' The deeper in the tree, the lower the value.
#' - *Gain* (for split nodes): Information gain metric of a split
#' (corresponds to the importance of the node in the model).
#' \item \code{Value} (for leafs): the margin value that the leaf may contribute to prediction.
#' }
#' The tree root nodes also indicate the Tree index (0-based).
#' - *Value* (for leaves): Margin value that the leaf may contribute to the prediction.
#'
#' The tree root nodes also indicate the tree index (0-based).
#'
#' The "Yes" branches are marked by the "< split_value" label.
#' The branches that also used for missing values are marked as bold
#' The branches also used for missing values are marked as bold
#' (as in "carrying extra capacity").
#'
#' This function uses \href{https://www.graphviz.org/}{GraphViz} as a backend of DiagrammeR.
#' This function uses [GraphViz](https://www.graphviz.org/) as DiagrammeR backend.
#'
#' @return
#'
#' When \code{render = TRUE}:
#' returns a rendered graph object which is an \code{htmlwidget} of class \code{grViz}.
#' Similar to ggplot objects, it needs to be printed to see it when not running from command line.
#'
#' When \code{render = FALSE}:
#' silently returns a graph object which is of DiagrammeR's class \code{dgr_graph}.
#' This could be useful if one wants to modify some of the graph attributes
#' before rendering the graph with \code{\link[DiagrammeR]{render_graph}}.
#' The value depends on the `render` parameter:
#' - If `render = TRUE` (default): Rendered graph object which is an htmlwidget of
#' class `grViz`. Similar to "ggplot" objects, it needs to be printed when not
#' running from the command line.
#' - If `render = FALSE`: Graph object which is of DiagrammeR's class `dgr_graph`.
#' This could be useful if one wants to modify some of the graph attributes
#' before rendering the graph with [DiagrammeR::render_graph()].
#'
#' @examples
#' data(agaricus.train, package='xgboost')
#' data(agaricus.train, package = "xgboost")
#'
#' bst <- xgboost(
#' data = agaricus.train$data,
#' label = agaricus.train$label,
#' max_depth = 3,
#' eta = 1,
#' nthread = 2,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#'
#' # plot the first tree, using the style from xgboost's core library
#' # (this plot should look identical to the ones generated from other
#' # interfaces like the python package for xgboost)
#' xgb.plot.tree(model = bst, trees = 1, style = "xgboost")
#'
#' bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 3,
#' eta = 1, nthread = 2, nrounds = 2,objective = "binary:logistic")
#' # plot all the trees
#' xgb.plot.tree(model = bst)
#' xgb.plot.tree(model = bst, trees = NULL)
#'
#' # plot only the first tree and display the node ID:
#' xgb.plot.tree(model = bst, trees = 0, show_node_id = TRUE)
#'
#' \dontrun{
#' # Below is an example of how to save this plot to a file.
#' # Note that for `export_graph` to work, the DiagrammeRsvg and rsvg packages must also be installed.
#' # Note that for export_graph() to work, the {DiagrammeRsvg}
#' # and {rsvg} packages must also be installed.
#'
#' library(DiagrammeR)
#' gr <- xgb.plot.tree(model=bst, trees=0:1, render=FALSE)
#' export_graph(gr, 'tree.pdf', width=1500, height=1900)
#' export_graph(gr, 'tree.png', width=1500, height=1900)
#'
#' gr <- xgb.plot.tree(model = bst, trees = 0:1, render = FALSE)
#' export_graph(gr, "tree.pdf", width = 1500, height = 1900)
#' export_graph(gr, "tree.png", width = 1500, height = 1900)
#' }
#'
#' @export
xgb.plot.tree <- function(feature_names = NULL, model = NULL, trees = NULL, plot_width = NULL, plot_height = NULL,
render = TRUE, show_node_id = FALSE, ...) {
render = TRUE, show_node_id = FALSE, style = c("R", "xgboost"), ...) {
check.deprecation(...)
if (!inherits(model, "xgb.Booster")) {
stop("model: Has to be an object of class xgb.Booster")
@ -78,9 +114,25 @@ xgb.plot.tree <- function(feature_names = NULL, model = NULL, trees = NULL, plot
stop("DiagrammeR package is required for xgb.plot.tree", call. = FALSE)
}
style <- as.character(head(style, 1L))
stopifnot(style %in% c("R", "xgboost"))
if (style == "xgboost") {
if (NROW(trees) != 1L || !render || show_node_id) {
stop("style='xgboost' is only supported for single, rendered tree, without node IDs.")
}
if (!is.null(feature_names)) {
stop(
"style='xgboost' cannot override 'feature_names'. Will automatically take them from the model."
)
}
txt <- xgb.dump(model, dump_format = "dot")
return(DiagrammeR::grViz(txt[[trees + 1]], width = plot_width, height = plot_height))
}
dt <- xgb.model.dt.tree(feature_names = feature_names, model = model, trees = trees)
dt[, label := paste0(Feature, "\nCover: ", Cover, ifelse(Feature == "Leaf", "\nValue: ", "\nGain: "), Quality)]
dt[, label := paste0(Feature, "\nCover: ", Cover, ifelse(Feature == "Leaf", "\nValue: ", "\nGain: "), Gain)]
if (show_node_id)
dt[, label := paste0(ID, ": ", label)]
dt[Node == 0, label := paste0("Tree ", Tree, "\n", label)]
@ -147,4 +199,4 @@ xgb.plot.tree <- function(feature_names = NULL, model = NULL, trees = NULL, plot
# Avoid error messages during CRAN check.
# The reason is that these variables are never declared
# They are mainly column names inferred by Data.table...
globalVariables(c("Feature", "ID", "Cover", "Quality", "Split", "Yes", "No", "Missing", ".", "shape", "filledcolor", "label"))
globalVariables(c("Feature", "ID", "Cover", "Gain", "Split", "Yes", "No", "Missing", ".", "shape", "filledcolor", "label"))

39
R-package/R/xgb.save.R
View File

@ -1,12 +1,24 @@
#' Save xgboost model to binary file
#'
#' Save xgboost model to a file in binary format.
#' Save xgboost model to a file in binary or JSON format.
#'
#' @param model model object of \code{xgb.Booster} class.
#' @param fname name of the file to write.
#' @param model Model object of \code{xgb.Booster} class.
#' @param fname Name of the file to write.
#'
#' Note that the extension of this file name determined the serialization format to use:\itemize{
#' \item Extension ".ubj" will use the universal binary JSON format (recommended).
#' This format uses binary types for e.g. floating point numbers, thereby preventing any loss
#' of precision when converting to a human-readable JSON text or similar.
#' \item Extension ".json" will use plain JSON, which is a human-readable format.
#' \item Extension ".deprecated" will use a \bold{deprecated} binary format. This format will
#' not be able to save attributes introduced after v1 of XGBoost, such as the "best_iteration"
#' attribute that boosters might keep, nor feature names or user-specifiec attributes.
#' \item If the format is not specified by passing one of the file extensions above, will
#' default to UBJ.
#' }
#'
#' @details
#' This methods allows to save a model in an xgboost-internal binary format which is universal
#' This methods allows to save a model in an xgboost-internal binary or text format which is universal
#' among the various xgboost interfaces. In R, the saved model file could be read-in later
#' using either the \code{\link{xgb.load}} function or the \code{xgb_model} parameter
#' of \code{\link{xgb.train}}.
@ -14,13 +26,13 @@
#' Note: a model can also be saved as an R-object (e.g., by using \code{\link[base]{readRDS}}
#' or \code{\link[base]{save}}). However, it would then only be compatible with R, and
#' corresponding R-methods would need to be used to load it. Moreover, persisting the model with
#' \code{\link[base]{readRDS}} or \code{\link[base]{save}}) will cause compatibility problems in
#' \code{\link[base]{readRDS}} or \code{\link[base]{save}}) might cause compatibility problems in
#' future versions of XGBoost. Consult \code{\link{a-compatibility-note-for-saveRDS-save}} to learn
#' how to persist models in a future-proof way, i.e. to make the model accessible in future
#' releases of XGBoost.
#'
#' @seealso
#' \code{\link{xgb.load}}, \code{\link{xgb.Booster.complete}}.
#' \code{\link{xgb.load}}
#'
#' @examples
#' data(agaricus.train, package='xgboost')
@ -32,15 +44,17 @@
#'
#' train <- agaricus.train
#' test <- agaricus.test
#' bst <- xgboost(
#' data = train$data, label = train$label, max_depth = 2, eta = 1,
#' bst <- xgb.train(
#' data = xgb.DMatrix(train$data, label = train$label),
#' max_depth = 2,
#' eta = 1,
#' nthread = nthread,
#' nrounds = 2,
#' objective = "binary:logistic"
#' )
#' xgb.save(bst, 'xgb.model')
#' bst <- xgb.load('xgb.model')
#' if (file.exists('xgb.model')) file.remove('xgb.model')
#' fname <- file.path(tempdir(), "xgb.ubj")
#' xgb.save(bst, fname)
#' bst <- xgb.load(fname)
#' @export
xgb.save <- function(model, fname) {
if (typeof(fname) != "character")
@ -49,8 +63,7 @@ xgb.save <- function(model, fname) {
stop("model must be xgb.Booster.",
if (inherits(model, "xgb.DMatrix")) " Use xgb.DMatrix.save to save an xgb.DMatrix object." else "")
}
model <- xgb.Booster.complete(model, saveraw = FALSE)
fname <- path.expand(fname)
.Call(XGBoosterSaveModel_R, model$handle, enc2utf8(fname[1]))
.Call(XGBoosterSaveModel_R, xgb.get.handle(model), enc2utf8(fname[1]))
return(TRUE)
}

8
R-package/R/xgb.save.raw.R
View File

@ -11,8 +11,6 @@
#' \item \code{deprecated}: Encode the booster into old customized binary format.
#' }
#'
#' Right now the default is \code{deprecated} but will be changed to \code{ubj} in upcoming release.
#'
#' @examples
#' data(agaricus.train, package='xgboost')
#' data(agaricus.test, package='xgboost')
@ -23,14 +21,14 @@
#'
#' train <- agaricus.train
#' test <- agaricus.test
#' bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
#' eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
#' bst <- xgb.train(data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
#' eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
#'
#' raw <- xgb.save.raw(bst)
#' bst <- xgb.load.raw(raw)
#'
#' @export
xgb.save.raw <- function(model, raw_format = "deprecated") {
xgb.save.raw <- function(model, raw_format = "ubj") {
handle <- xgb.get.handle(model)
args <- list(format = raw_format)
.Call(XGBoosterSaveModelToRaw_R, handle, jsonlite::toJSON(args, auto_unbox = TRUE))

21
R-package/R/xgb.serialize.R
View File

@ -1,21 +0,0 @@
#' Serialize the booster instance into R's raw vector. The serialization method differs
#' from \code{\link{xgb.save.raw}} as the latter one saves only the model but not
#' parameters. This serialization format is not stable across different xgboost versions.
#'
#' @param booster the booster instance
#'
#' @examples
#' data(agaricus.train, package='xgboost')
#' data(agaricus.test, package='xgboost')
#' train <- agaricus.train
#' test <- agaricus.test
#' bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
#' eta = 1, nthread = 2, nrounds = 2,objective = "binary:logistic")
#' raw <- xgb.serialize(bst)
#' bst <- xgb.unserialize(raw)
#'
#' @export
xgb.serialize <- function(booster) {
handle <- xgb.get.handle(booster)
.Call(XGBoosterSerializeToBuffer_R, handle)
}

130
R-package/R/xgb.train.R
View File

@ -152,6 +152,10 @@
#' See \code{\link{callbacks}}. Some of the callbacks are automatically created depending on the
#' parameters' values. User can provide either existing or their own callback methods in order
#' to customize the training process.
#'
#' Note that some callbacks might try to set an evaluation log - be aware that these evaluation logs
#' are kept as R attributes, and thus do not get saved when using non-R serializaters like
#' \link{xgb.save} (but are kept when using R serializers like \link{saveRDS}).
#' @param ... other parameters to pass to \code{params}.
#' @param label vector of response values. Should not be provided when data is
#' a local data file name or an \code{xgb.DMatrix}.
@ -160,6 +164,9 @@
#' This parameter is only used when input is a dense matrix.
#' @param weight a vector indicating the weight for each row of the input.
#'
#' @return
#' An object of class \code{xgb.Booster}.
#'
#' @details
#' These are the training functions for \code{xgboost}.
#'
@ -201,28 +208,20 @@
#' \item \code{cb.save.model}: when \code{save_period > 0} is set.
#' }
#'
#' @return
#' An object of class \code{xgb.Booster} with the following elements:
#' \itemize{
#' \item \code{handle} a handle (pointer) to the xgboost model in memory.
#' \item \code{raw} a cached memory dump of the xgboost model saved as R's \code{raw} type.
#' \item \code{niter} number of boosting iterations.
#' \item \code{evaluation_log} evaluation history stored as a \code{data.table} with the
#' first column corresponding to iteration number and the rest corresponding to evaluation
#' metrics' values. It is created by the \code{\link{cb.evaluation.log}} callback.
#' \item \code{call} a function call.
#' \item \code{params} parameters that were passed to the xgboost library. Note that it does not
#' capture parameters changed by the \code{\link{cb.reset.parameters}} callback.
#' \item \code{callbacks} callback functions that were either automatically assigned or
#' explicitly passed.
#' \item \code{best_iteration} iteration number with the best evaluation metric value
#' (only available with early stopping).
#' \item \code{best_score} the best evaluation metric value during early stopping.
#' (only available with early stopping).
#' \item \code{feature_names} names of the training dataset features
#' (only when column names were defined in training data).
#' \item \code{nfeatures} number of features in training data.
#' }
#' Note that objects of type `xgb.Booster` as returned by this function behave a bit differently
#' from typical R objects (it's an 'altrep' list class), and it makes a separation between
#' internal booster attributes (restricted to jsonifyable data), accessed through \link{xgb.attr}
#' and shared between interfaces through serialization functions like \link{xgb.save}; and
#' R-specific attributes, accessed through \link{attributes} and \link{attr}, which are otherwise
#' only used in the R interface, only kept when using R's serializers like \link{saveRDS}, and
#' not anyhow used by functions like \link{predict.xgb.Booster}.
#'
#' Be aware that one such R attribute that is automatically added is `params` - this attribute
#' is assigned from the `params` argument to this function, and is only meant to serve as a
#' reference for what went into the booster, but is not used in other methods that take a booster
#' object - so for example, changing the booster's configuration requires calling `xgb.config<-`
#' or 'xgb.parameters<-', while simply modifying `attributes(model)$params$<...>` will have no
#' effect elsewhere.
#'
#' @seealso
#' \code{\link{callbacks}},
@ -251,9 +250,9 @@
#' watchlist <- list(train = dtrain, eval = dtest)
#'
#' ## A simple xgb.train example:
#' param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
#' param <- list(max_depth = 2, eta = 1, nthread = nthread,
#' objective = "binary:logistic", eval_metric = "auc")
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist)
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0)
#'
#' ## An xgb.train example where custom objective and evaluation metric are
#' ## used:
@ -272,13 +271,13 @@
#'
#' # These functions could be used by passing them either:
#' # as 'objective' and 'eval_metric' parameters in the params list:
#' param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
#' param <- list(max_depth = 2, eta = 1, nthread = nthread,
#' objective = logregobj, eval_metric = evalerror)
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist)
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0)
#'
#' # or through the ... arguments:
#' param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread)
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist,
#' param <- list(max_depth = 2, eta = 1, nthread = nthread)
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
#' objective = logregobj, eval_metric = evalerror)
#'
#' # or as dedicated 'obj' and 'feval' parameters of xgb.train:
@ -287,10 +286,10 @@
#'
#'
#' ## An xgb.train example of using variable learning rates at each iteration:
#' param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
#' param <- list(max_depth = 2, eta = 1, nthread = nthread,
#' objective = "binary:logistic", eval_metric = "auc")
#' my_etas <- list(eta = c(0.5, 0.1))
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist,
#' bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
#' callbacks = list(cb.reset.parameters(my_etas)))
#'
#' ## Early stopping:
@ -371,27 +370,31 @@ xgb.train <- function(params = list(), data, nrounds, watchlist = list(),
# The tree updating process would need slightly different handling
is_update <- NVL(params[['process_type']], '.') == 'update'
past_evaluation_log <- NULL
if (inherits(xgb_model, "xgb.Booster")) {
past_evaluation_log <- attributes(xgb_model)$evaluation_log
}
# Construct a booster (either a new one or load from xgb_model)
handle <- xgb.Booster.handle(
bst <- xgb.Booster(
params = params,
cachelist = append(watchlist, dtrain),
modelfile = xgb_model,
handle = NULL
modelfile = xgb_model
)
niter_init <- bst$niter
bst <- bst$bst
.Call(
XGBoosterCopyInfoFromDMatrix_R,
xgb.get.handle(bst),
dtrain
)
bst <- xgb.handleToBooster(handle = handle, raw = NULL)
# extract parameters that can affect the relationship b/w #trees and #iterations
num_class <- max(as.numeric(NVL(params[['num_class']], 1)), 1)
num_parallel_tree <- max(as.numeric(NVL(params[['num_parallel_tree']], 1)), 1)
# Note: it might look like these aren't used, but they need to be defined in this
# environment for the callbacks for work correctly.
num_class <- max(as.numeric(NVL(params[['num_class']], 1)), 1) # nolint
num_parallel_tree <- max(as.numeric(NVL(params[['num_parallel_tree']], 1)), 1) # nolint
# When the 'xgb_model' was set, find out how many boosting iterations it has
niter_init <- 0
if (!is.null(xgb_model)) {
niter_init <- as.numeric(xgb.attr(bst, 'niter')) + 1
if (length(niter_init) == 0) {
niter_init <- xgb.ntree(bst) %/% (num_parallel_tree * num_class)
}
}
if (is_update && nrounds > niter_init)
stop("nrounds cannot be larger than ", niter_init, " (nrounds of xgb_model)")
@ -405,7 +408,7 @@ xgb.train <- function(params = list(), data, nrounds, watchlist = list(),
for (f in cb$pre_iter) f()
xgb.iter.update(
booster_handle = bst$handle,
bst = bst,
dtrain = dtrain,
iter = iteration - 1,
obj = obj
@ -413,46 +416,43 @@ xgb.train <- function(params = list(), data, nrounds, watchlist = list(),
if (length(watchlist) > 0) {
bst_evaluation <- xgb.iter.eval( # nolint: object_usage_linter
booster_handle = bst$handle,
bst = bst,
watchlist = watchlist,
iter = iteration - 1,
feval = feval
)
}
xgb.attr(bst$handle, 'niter') <- iteration - 1
for (f in cb$post_iter) f()
if (stop_condition) break
}
for (f in cb$finalize) f(finalize = TRUE)
bst <- xgb.Booster.complete(bst, saveraw = TRUE)
# store the total number of boosting iterations
bst$niter <- end_iteration
# store the evaluation results
if (length(evaluation_log) > 0 &&
nrow(evaluation_log) > 0) {
keep_evaluation_log <- FALSE
if (length(evaluation_log) > 0 && nrow(evaluation_log) > 0) {
keep_evaluation_log <- TRUE
# include the previous compatible history when available
if (inherits(xgb_model, 'xgb.Booster') &&
!is_update &&
!is.null(xgb_model$evaluation_log) &&
!is.null(past_evaluation_log) &&
isTRUE(all.equal(colnames(evaluation_log),
colnames(xgb_model$evaluation_log)))) {
evaluation_log <- rbindlist(list(xgb_model$evaluation_log, evaluation_log))
colnames(past_evaluation_log)))) {
evaluation_log <- rbindlist(list(past_evaluation_log, evaluation_log))
}
bst$evaluation_log <- evaluation_log
}
bst$call <- match.call()
bst$params <- params
bst$callbacks <- callbacks
if (!is.null(colnames(dtrain)))
bst$feature_names <- colnames(dtrain)
bst$nfeatures <- ncol(dtrain)
extra_attrs <- list(
call = match.call(),
params = params,
callbacks = callbacks
)
if (keep_evaluation_log) {
extra_attrs$evaluation_log <- evaluation_log
}
curr_attrs <- attributes(bst)
attributes(bst) <- c(curr_attrs, extra_attrs)
return(bst)
}

41
R-package/R/xgb.unserialize.R
View File

@ -1,41 +0,0 @@
#' Load the instance back from \code{\link{xgb.serialize}}
#'
#' @param buffer the buffer containing booster instance saved by \code{\link{xgb.serialize}}
#' @param handle An \code{xgb.Booster.handle} object which will be overwritten with
#' the new deserialized object. Must be a null handle (e.g. when loading the model through
#' `readRDS`). If not provided, a new handle will be created.
#' @return An \code{xgb.Booster.handle} object.
#'
#' @export
xgb.unserialize <- function(buffer, handle = NULL) {
cachelist <- list()
if (is.null(handle)) {
handle <- .Call(XGBoosterCreate_R, cachelist)
} else {
if (!is.null.handle(handle))
stop("'handle' is not null/empty. Cannot overwrite existing handle.")
.Call(XGBoosterCreateInEmptyObj_R, cachelist, handle)
}
tryCatch(
.Call(XGBoosterUnserializeFromBuffer_R, handle, buffer),
error = function(e) {
error_msg <- conditionMessage(e)
m <- regexec("(src[\\\\/]learner.cc:[0-9]+): Check failed: (header == serialisation_header_)",
error_msg, perl = TRUE)
groups <- regmatches(error_msg, m)[[1]]
if (length(groups) == 3) {
warning(paste("The model had been generated by XGBoost version 1.0.0 or earlier and was ",
"loaded from a RDS file. We strongly ADVISE AGAINST using saveRDS() ",
"function, to ensure that your model can be read in current and upcoming ",
"XGBoost releases. Please use xgb.save() instead to preserve models for the ",
"long term. For more details and explanation, see ",
"https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html",
sep = ""))
.Call(XGBoosterLoadModelFromRaw_R, handle, buffer)
} else {
stop(e)
}
})
class(handle) <- "xgb.Booster.handle"
return(handle)
}

13
R-package/R/xgboost.R
View File

@ -40,10 +40,10 @@ xgboost <- function(data = NULL, label = NULL, missing = NA, weight = NULL,
#' }
#'
#' @references
#' https://archive.ics.uci.edu/ml/datasets/Mushroom
#' <https://archive.ics.uci.edu/ml/datasets/Mushroom>
#'
#' Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository
#' [http://archive.ics.uci.edu/ml]. Irvine, CA: University of California,
#' <http://archive.ics.uci.edu/ml>. Irvine, CA: University of California,
#' School of Information and Computer Science.
#'
#' @docType data
@ -67,10 +67,10 @@ NULL
#' }
#'
#' @references
#' https://archive.ics.uci.edu/ml/datasets/Mushroom
#' <https://archive.ics.uci.edu/ml/datasets/Mushroom>
#'
#' Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository
#' [http://archive.ics.uci.edu/ml]. Irvine, CA: University of California,
#' <http://archive.ics.uci.edu/ml>. Irvine, CA: University of California,
#' School of Information and Computer Science.
#'
#' @docType data
@ -82,7 +82,7 @@ NULL
NULL
# Various imports
#' @importClassesFrom Matrix dgCMatrix dgeMatrix
#' @importClassesFrom Matrix dgCMatrix dgeMatrix dgRMatrix
#' @importFrom Matrix colSums
#' @importFrom Matrix sparse.model.matrix
#' @importFrom Matrix sparseVector
@ -98,9 +98,12 @@ NULL
#' @importFrom data.table setnames
#' @importFrom jsonlite fromJSON
#' @importFrom jsonlite toJSON
#' @importFrom methods new
#' @importFrom utils object.size str tail
#' @importFrom stats coef
#' @importFrom stats predict
#' @importFrom stats median
#' @importFrom stats variable.names
#' @importFrom utils head
#' @importFrom graphics barplot
#' @importFrom graphics lines

1
R-package/demo/00Index
View File

@ -1,5 +1,4 @@
basic_walkthrough Basic feature walkthrough
caret_wrapper Use xgboost to train in caret library
custom_objective Customize loss function, and evaluation metric
boost_from_prediction Boosting from existing prediction
predict_first_ntree Predicting using first n trees

1
R-package/demo/README.md
View File

@ -1,7 +1,6 @@
XGBoost R Feature Walkthrough
====
* [Basic walkthrough of wrappers](basic_walkthrough.R)
* [Train a xgboost model from caret library](caret_wrapper.R)
* [Customize loss function, and evaluation metric](custom_objective.R)
* [Boosting from existing prediction](boost_from_prediction.R)
* [Predicting using first n trees](predict_first_ntree.R)

44
R-package/demo/caret_wrapper.R
View File

@ -1,44 +0,0 @@
# install development version of caret library that contains xgboost models
require(caret)
require(xgboost)
require(data.table)
require(vcd)
require(e1071)
# Load Arthritis dataset in memory.
data(Arthritis)
# Create a copy of the dataset with data.table package
# (data.table is 100% compliant with R dataframe but its syntax is a lot more consistent
# and its performance are really good).
df <- data.table(Arthritis, keep.rownames = FALSE)
# Let's add some new categorical features to see if it helps.
# Of course these feature are highly correlated to the Age feature.
# Usually it's not a good thing in ML, but Tree algorithms (including boosted trees) are able to select the best features,
# even in case of highly correlated features.
# For the first feature we create groups of age by rounding the real age.
# Note that we transform it to factor (categorical data) so the algorithm treat them as independant values.
df[, AgeDiscret := as.factor(round(Age / 10, 0))]
# Here is an even stronger simplification of the real age with an arbitrary split at 30 years old.
# I choose this value based on nothing.
# We will see later if simplifying the information based on arbitrary values is a good strategy
# (I am sure you already have an idea of how well it will work!).
df[, AgeCat := as.factor(ifelse(Age > 30, "Old", "Young"))]
# We remove ID as there is nothing to learn from this feature (it will just add some noise as the dataset is small).
df[, ID := NULL]
#-------------Basic Training using XGBoost in caret Library-----------------
# Set up control parameters for caret::train
# Here we use 10-fold cross-validation, repeating twice, and using random search for tuning hyper-parameters.
fitControl <- trainControl(method = "repeatedcv", number = 10, repeats = 2, search = "random")
# train a xgbTree model using caret::train
model <- train(factor(Improved) ~ ., data = df, method = "xgbTree", trControl = fitControl)
# Instead of tree for our boosters, you can also fit a linear regression or logistic regression model
# using xgbLinear
# model <- train(factor(Improved)~., data = df, method = "xgbLinear", trControl = fitControl)
# See model results
print(model)

4
R-package/demo/create_sparse_matrix.R
View File

@ -81,8 +81,8 @@ output_vector <- df[, Y := 0][Improved == "Marked", Y := 1][, Y]
# Following is the same process as other demo
cat("Learning...\n")
bst <- xgboost(data = sparse_matrix, label = output_vector, max_depth = 9,
eta = 1, nthread = 2, nrounds = 10, objective = "binary:logistic")
bst <- xgb.train(data = xgb.DMatrix(sparse_matrix, label = output_vector), max_depth = 9,
eta = 1, nthread = 2, nrounds = 10, objective = "binary:logistic")
importance <- xgb.importance(feature_names = colnames(sparse_matrix), model = bst)
print(importance)

18
R-package/demo/interaction_constraints.R
View File

@ -74,26 +74,26 @@ cols2ids <- function(object, col_names) {
interaction_list_fid <- cols2ids(interaction_list, colnames(train))
# Fit model with interaction constraints
bst <- xgboost(data = train, label = y, max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000,
interaction_constraints = interaction_list_fid)
bst <- xgb.train(data = xgb.DMatrix(train, label = y), max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000,
interaction_constraints = interaction_list_fid)
bst_tree <- xgb.model.dt.tree(colnames(train), bst)
bst_interactions <- treeInteractions(bst_tree, 4)
# interactions constrained to combinations of V1*V2 and V3*V4*V5
# Fit model without interaction constraints
bst2 <- xgboost(data = train, label = y, max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000)
bst2 <- xgb.train(data = xgb.DMatrix(train, label = y), max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000)
bst2_tree <- xgb.model.dt.tree(colnames(train), bst2)
bst2_interactions <- treeInteractions(bst2_tree, 4) # much more interactions
# Fit model with both interaction and monotonicity constraints
bst3 <- xgboost(data = train, label = y, max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000,
interaction_constraints = interaction_list_fid,
monotone_constraints = c(-1, 0, 0, 0, 0, 0, 0, 0, 0, 0))
bst3 <- xgb.train(data = xgb.DMatrix(train, label = y), max_depth = 4,
eta = 0.1, nthread = 2, nrounds = 1000,
interaction_constraints = interaction_list_fid,
monotone_constraints = c(-1, 0, 0, 0, 0, 0, 0, 0, 0, 0))
bst3_tree <- xgb.model.dt.tree(colnames(train), bst3)
bst3_interactions <- treeInteractions(bst3_tree, 4)

4
R-package/demo/poisson_regression.R
View File

@ -1,6 +1,6 @@
data(mtcars)
head(mtcars)
bst <- xgboost(data = as.matrix(mtcars[, -11]), label = mtcars[, 11],
objective = 'count:poisson', nrounds = 5)
bst <- xgb.train(data = xgb.DMatrix(as.matrix(mtcars[, -11]), label = mtcars[, 11]),
objective = 'count:poisson', nrounds = 5)
pred <- predict(bst, as.matrix(mtcars[, -11]))
sqrt(mean((pred - mtcars[, 11]) ^ 2))

2
R-package/demo/predict_leaf_indices.R
View File

@ -27,7 +27,7 @@ head(pred_with_leaf)
create.new.tree.features <- function(model, original.features) {
pred_with_leaf <- predict(model, original.features, predleaf = TRUE)
cols <- list()
for (i in 1:model$niter) {
for (i in 1:xgb.get.num.boosted.rounds(model)) {
# max is not the real max but it s not important for the purpose of adding features
leaf.id <- sort(unique(pred_with_leaf[, i]))
cols[[i]] <- factor(x = pred_with_leaf[, i], level = leaf.id)

1
R-package/demo/runall.R
View File

@ -9,6 +9,5 @@ demo(create_sparse_matrix, package = 'xgboost')
demo(predict_leaf_indices, package = 'xgboost')
demo(early_stopping, package = 'xgboost')
demo(poisson_regression, package = 'xgboost')
demo(caret_wrapper, package = 'xgboost')
demo(tweedie_regression, package = 'xgboost')
#demo(gpu_accelerated, package = 'xgboost') # can only run when built with GPU support

2
R-package/inst/make-r-def.R
View File

@ -55,7 +55,7 @@ message(sprintf("Creating '%s' from '%s'", OUT_DEF_FILE, IN_DLL_FILE))
}
# use objdump to dump all the symbols
OBJDUMP_FILE <- "objdump-out.txt"
OBJDUMP_FILE <- file.path(tempdir(), "objdump-out.txt")
.pipe_shell_command_to_stdout(
command = "objdump"
, args = c("-p", IN_DLL_FILE)

75
R-package/man/a-compatibility-note-for-saveRDS-save.Rd
View File

@ -2,16 +2,44 @@
% Please edit documentation in R/utils.R
\name{a-compatibility-note-for-saveRDS-save}
\alias{a-compatibility-note-for-saveRDS-save}
\title{Do not use \code{\link[base]{saveRDS}} or \code{\link[base]{save}} for long-term archival of
models. Instead, use \code{\link{xgb.save}} or \code{\link{xgb.save.raw}}.}
\title{Model Serialization and Compatibility}
\description{
It is a common practice to use the built-in \code{\link[base]{saveRDS}} function (or
\code{\link[base]{save}}) to persist R objects to the disk. While it is possible to persist
\code{xgb.Booster} objects using \code{\link[base]{saveRDS}}, it is not advisable to do so if
the model is to be accessed in the future. If you train a model with the current version of
XGBoost and persist it with \code{\link[base]{saveRDS}}, the model is not guaranteed to be
accessible in later releases of XGBoost. To ensure that your model can be accessed in future
releases of XGBoost, use \code{\link{xgb.save}} or \code{\link{xgb.save.raw}} instead.
When it comes to serializing XGBoost models, it's possible to use R serializers such as
\link{save} or \link{saveRDS} to serialize an XGBoost R model, but XGBoost also provides
its own serializers with better compatibility guarantees, which allow loading
said models in other language bindings of XGBoost.
Note that an \code{xgb.Booster} object, outside of its core components, might also keep:\itemize{
\item Additional model configuration (accessible through \link{xgb.config}),
which includes model fitting parameters like \code{max_depth} and runtime parameters like \code{nthread}.
These are not necessarily useful for prediction/importance/plotting.
\item Additional R-specific attributes - e.g. results of callbacks, such as evaluation logs,
which are kept as a \code{data.table} object, accessible through \code{attributes(model)$evaluation_log}
if present.
}
The first one (configurations) does not have the same compatibility guarantees as
the model itself, including attributes that are set and accessed through \link{xgb.attributes} - that is, such configuration
might be lost after loading the booster in a different XGBoost version, regardless of the
serializer that was used. These are saved when using \link{saveRDS}, but will be discarded
if loaded into an incompatible XGBoost version. They are not saved when using XGBoost's
serializers from its public interface including \link{xgb.save} and \link{xgb.save.raw}.
The second ones (R attributes) are not part of the standard XGBoost model structure, and thus are
not saved when using XGBoost's own serializers. These attributes are only used for informational
purposes, such as keeping track of evaluation metrics as the model was fit, or saving the R
call that produced the model, but are otherwise not used for prediction / importance / plotting / etc.
These R attributes are only preserved when using R's serializers.
Note that XGBoost models in R starting from version \verb{2.1.0} and onwards, and XGBoost models
before version \verb{2.1.0}; have a very different R object structure and are incompatible with
each other. Hence, models that were saved with R serializers live \code{saveRDS} or \code{save} before
version \verb{2.1.0} will not work with latter \code{xgboost} versions and vice versa. Be aware that
the structure of R model objects could in theory change again in the future, so XGBoost's serializers
should be preferred for long-term storage.
Furthermore, note that using the package \code{qs} for serialization will require version 0.26 or
higher of said package, and will have the same compatibility restrictions as R serializers.
}
\details{
Use \code{\link{xgb.save}} to save the XGBoost model as a stand-alone file. You may opt into
@ -24,26 +52,29 @@ re-construct the corresponding model. To read the model back, use \code{\link{xg
The \code{\link{xgb.save.raw}} function is useful if you'd like to persist the XGBoost model
as part of another R object.
Note: Do not use \code{\link{xgb.serialize}} to store models long-term. It persists not only the
model but also internal configurations and parameters, and its format is not stable across
multiple XGBoost versions. Use \code{\link{xgb.serialize}} only for checkpointing.
Use \link{saveRDS} if you require the R-specific attributes that a booster might have, such
as evaluation logs, but note that future compatibility of such objects is outside XGBoost's
control as it relies on R's serialization format (see e.g. the details section in
\link{serialize} and \link{save} from base R).
For more details and explanation about model persistence and archival, consult the page
\url{https://xgboost.readthedocs.io/en/latest/tutorials/saving_model.html}.
}
\examples{
data(agaricus.train, package='xgboost')
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
bst <- xgb.train(data = xgb.DMatrix(agaricus.train$data, label = agaricus.train$label),
max_depth = 2, eta = 1, nthread = 2, nrounds = 2,
objective = "binary:logistic")
# Save as a stand-alone file; load it with xgb.load()
xgb.save(bst, 'xgb.model')
bst2 <- xgb.load('xgb.model')
fname <- file.path(tempdir(), "xgb_model.ubj")
xgb.save(bst, fname)
bst2 <- xgb.load(fname)
# Save as a stand-alone file (JSON); load it with xgb.load()
xgb.save(bst, 'xgb.model.json')
bst2 <- xgb.load('xgb.model.json')
if (file.exists('xgb.model.json')) file.remove('xgb.model.json')
fname <- file.path(tempdir(), "xgb_model.json")
xgb.save(bst, fname)
bst2 <- xgb.load(fname)
# Save as a raw byte vector; load it with xgb.load.raw()
xgb_bytes <- xgb.save.raw(bst)
@ -54,11 +85,11 @@ obj <- list(xgb_model_bytes = xgb.save.raw(bst), description = "My first XGBoost
# Persist the R object. Here, saveRDS() is okay, since it doesn't persist
# xgb.Booster directly. What's being persisted is the future-proof byte representation
# as given by xgb.save.raw().
saveRDS(obj, 'my_object.rds')
fname <- file.path(tempdir(), "my_object.Rds")
saveRDS(obj, fname)
# Read back the R object
obj2 <- readRDS('my_object.rds')
obj2 <- readRDS(fname)
# Re-construct xgb.Booster object from the bytes
bst2 <- xgb.load.raw(obj2$xgb_model_bytes)
if (file.exists('my_object.rds')) file.remove('my_object.rds')
}

8
R-package/man/agaricus.test.Rd
View File

@ -19,15 +19,15 @@ UCI Machine Learning Repository.
This data set includes the following fields:
\itemize{
\item \code{label} the label for each record
\item \code{data} a sparse Matrix of \code{dgCMatrix} class, with 126 columns.
\item \code{label} the label for each record
\item \code{data} a sparse Matrix of \code{dgCMatrix} class, with 126 columns.
}
}
\references{
https://archive.ics.uci.edu/ml/datasets/Mushroom
\url{https://archive.ics.uci.edu/ml/datasets/Mushroom}
Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository
[http://archive.ics.uci.edu/ml]. Irvine, CA: University of California,
\url{http://archive.ics.uci.edu/ml}. Irvine, CA: University of California,
School of Information and Computer Science.
}
\keyword{datasets}

8
R-package/man/agaricus.train.Rd
View File

@ -19,15 +19,15 @@ UCI Machine Learning Repository.
This data set includes the following fields:
\itemize{
\item \code{label} the label for each record
\item \code{data} a sparse Matrix of \code{dgCMatrix} class, with 126 columns.
\item \code{label} the label for each record
\item \code{data} a sparse Matrix of \code{dgCMatrix} class, with 126 columns.
}
}
\references{
https://archive.ics.uci.edu/ml/datasets/Mushroom
\url{https://archive.ics.uci.edu/ml/datasets/Mushroom}
Bache, K. & Lichman, M. (2013). UCI Machine Learning Repository
[http://archive.ics.uci.edu/ml]. Irvine, CA: University of California,
\url{http://archive.ics.uci.edu/ml}. Irvine, CA: University of California,
School of Information and Computer Science.
}
\keyword{datasets}

17
R-package/man/cb.save.model.Rd
View File

@ -4,17 +4,22 @@
\alias{cb.save.model}
\title{Callback closure for saving a model file.}
\usage{
cb.save.model(save_period = 0, save_name = "xgboost.model")
cb.save.model(save_period = 0, save_name = "xgboost.ubj")
}
\arguments{
\item{save_period}{save the model to disk after every
\code{save_period} iterations; 0 means save the model at the end.}
\item{save_name}{the name or path for the saved model file.
It can contain a \code{\link[base]{sprintf}} formatting specifier
to include the integer iteration number in the file name.
E.g., with \code{save_name} = 'xgboost_%04d.model',
the file saved at iteration 50 would be named "xgboost_0050.model".}
\if{html}{\out{<div class="sourceCode">}}\preformatted{ Note that the format of the model being saved is determined by the file
extension specified here (see \link{xgb.save} for details about how it works).
It can contain a \code{\link[base]{sprintf}} formatting specifier
to include the integer iteration number in the file name.
E.g., with \code{save_name} = 'xgboost_\%04d.ubj',
the file saved at iteration 50 would be named "xgboost_0050.ubj".
}\if{html}{\out{</div>}}}
}
\description{
Callback closure for saving a model file.
@ -29,5 +34,7 @@ Callback function expects the following values to be set in its calling frame:
\code{end_iteration}.
}
\seealso{
\link{xgb.save}
\code{\link{callbacks}}
}

50
R-package/man/coef.xgb.Booster.Rd Normal file
View File

@ -0,0 +1,50 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{coef.xgb.Booster}
\alias{coef.xgb.Booster}
\title{Extract coefficients from linear booster}
\usage{
\method{coef}{xgb.Booster}(object, ...)
}
\arguments{
\item{object}{A fitted booster of 'gblinear' type.}
\item{...}{Not used.}
}
\value{
The extracted coefficients:\itemize{
\item If there's only one coefficient per column in the data, will be returned as a
vector, potentially containing the feature names if available, with the intercept
as first column.
\item If there's more than one coefficient per column in the data (e.g. when using
\code{objective="multi:softmax"}), will be returned as a matrix with dimensions equal
to \verb{[num_features, num_cols]}, with the intercepts as first row. Note that the column
(classes in multi-class classification) dimension will not be named.
}
The intercept returned here will include the 'base_score' parameter (unlike the 'bias'
or the last coefficient in the model dump, which doesn't have 'base_score' added to it),
hence one should get the same values from calling \code{predict(..., outputmargin = TRUE)} and
from performing a matrix multiplication with \code{model.matrix(~., ...)}.
Be aware that the coefficients are obtained by first converting them to strings and
back, so there will always be some very small lose of precision compared to the actual
coefficients as used by \link{predict.xgb.Booster}.
}
\description{
Extracts the coefficients from a 'gblinear' booster object,
as produced by \code{xgb.train} when using parameter \code{booster="gblinear"}.
Note: this function will error out if passing a booster model
which is not of "gblinear" type.
}
\examples{
library(xgboost)
data(mtcars)
y <- mtcars[, 1]
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(data = x, label = y, nthread = 1)
params <- list(booster = "gblinear", nthread = 1)
model <- xgb.train(data = dm, params = params, nrounds = 2)
coef(model)
}

79
R-package/man/getinfo.Rd
View File

@ -1,36 +1,78 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{getinfo}
% Please edit documentation in R/xgb.Booster.R, R/xgb.DMatrix.R
\name{getinfo.xgb.Booster}
\alias{getinfo.xgb.Booster}
\alias{setinfo.xgb.Booster}
\alias{getinfo}
\alias{getinfo.xgb.DMatrix}
\title{Get information of an xgb.DMatrix object}
\alias{setinfo}
\alias{setinfo.xgb.DMatrix}
\title{Get or set information of xgb.DMatrix and xgb.Booster objects}
\usage{
getinfo(object, ...)
\method{getinfo}{xgb.Booster}(object, name)
\method{getinfo}{xgb.DMatrix}(object, name, ...)
\method{setinfo}{xgb.Booster}(object, name, info)
getinfo(object, name)
\method{getinfo}{xgb.DMatrix}(object, name)
setinfo(object, name, info)
\method{setinfo}{xgb.DMatrix}(object, name, info)
}
\arguments{
\item{object}{Object of class \code{xgb.DMatrix}}
\item{...}{other parameters}
\item{object}{Object of class \code{xgb.DMatrix} of \code{xgb.Booster}.}
\item{name}{the name of the information field to get (see details)}
\item{info}{the specific field of information to set}
}
\value{
For \code{getinfo}, will return the requested field. For \code{setinfo}, will always return value \code{TRUE}
if it succeeds.
}
\description{
Get information of an xgb.DMatrix object
Get or set information of xgb.DMatrix and xgb.Booster objects
}
\details{
The \code{name} field can be one of the following:
The \code{name} field can be one of the following for \code{xgb.DMatrix}:
\itemize{
\item \code{label}: label XGBoost learn from ;
\item \code{weight}: to do a weight rescale ;
\item \code{base_margin}: base margin is the base prediction XGBoost will boost from ;
\item \code{nrow}: number of rows of the \code{xgb.DMatrix}.
\item \code{label}
\item \code{weight}
\item \code{base_margin}
\item \code{label_lower_bound}
\item \code{label_upper_bound}
\item \code{group}
\item \code{feature_type}
\item \code{feature_name}
\item \code{nrow}
}
See the documentation for \link{xgb.DMatrix} for more information about these fields.
For \code{xgb.Booster}, can be one of the following:
\itemize{
\item \code{feature_type}
\item \code{feature_name}
}
\code{group} can be setup by \code{setinfo} but can't be retrieved by \code{getinfo}.
Note that, while 'qid' cannot be retrieved, it's possible to get the equivalent 'group'
for a DMatrix that had 'qid' assigned.
\bold{Important}: when calling \code{setinfo}, the objects are modified in-place. See
\link{xgb.copy.Booster} for an idea of this in-place assignment works.
See the documentation for \link{xgb.DMatrix} for possible fields that can be set
(which correspond to arguments in that function).
Note that the following fields are allowed in the construction of an \code{xgb.DMatrix}
but \bold{aren't} allowed here:\itemize{
\item data
\item missing
\item silent
\item nthread
}
}
\examples{
data(agaricus.train, package='xgboost')
@ -41,4 +83,11 @@ setinfo(dtrain, 'label', 1-labels)
labels2 <- getinfo(dtrain, 'label')
stopifnot(all(labels2 == 1-labels))
data(agaricus.train, package='xgboost')
dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
labels <- getinfo(dtrain, 'label')
setinfo(dtrain, 'label', 1-labels)
labels2 <- getinfo(dtrain, 'label')
stopifnot(all.equal(labels2, 1-labels))
}

186
R-package/man/predict.xgb.Booster.Rd
View File

@ -2,8 +2,7 @@
% Please edit documentation in R/xgb.Booster.R
\name{predict.xgb.Booster}
\alias{predict.xgb.Booster}
\alias{predict.xgb.Booster.handle}
\title{Predict method for eXtreme Gradient Boosting model}
\title{Predict method for XGBoost model}
\usage{
\method{predict}{xgb.Booster}(
object,
@ -21,94 +20,88 @@
strict_shape = FALSE,
...
)
\method{predict}{xgb.Booster.handle}(object, ...)
}
\arguments{
\item{object}{Object of class \code{xgb.Booster} or \code{xgb.Booster.handle}}
\item{object}{Object of class \code{xgb.Booster}.}
\item{newdata}{takes \code{matrix}, \code{dgCMatrix}, \code{dgRMatrix}, \code{dsparseVector},
local data file or \code{xgb.DMatrix}.
\item{newdata}{Takes \code{matrix}, \code{dgCMatrix}, \code{dgRMatrix}, \code{dsparseVector},
local data file, or \code{xgb.DMatrix}.
For single-row predictions on sparse data, it is recommended to use the CSR format.
If passing a sparse vector, it will take it as a row vector.}
For single-row predictions on sparse data, it's recommended to use CSR format. If passing
a sparse vector, it will take it as a row vector.}
\item{missing}{Only used when input is a dense matrix. Pick a float value that represents
missing values in data (e.g., 0 or some other extreme value).}
\item{missing}{Missing is only used when input is dense matrix. Pick a float value that represents
missing values in data (e.g., sometimes 0 or some other extreme value is used).}
\item{outputmargin}{whether the prediction should be returned in the for of original untransformed
\item{outputmargin}{Whether the prediction should be returned in the form of original untransformed
sum of predictions from boosting iterations' results. E.g., setting \code{outputmargin=TRUE} for
logistic regression would result in predictions for log-odds instead of probabilities.}
logistic regression would return log-odds instead of probabilities.}
\item{ntreelimit}{Deprecated, use \code{iterationrange} instead.}
\item{predleaf}{whether predict leaf index.}
\item{predleaf}{Whether to predict pre-tree leaf indices.}
\item{predcontrib}{whether to return feature contributions to individual predictions (see Details).}
\item{predcontrib}{Whether to return feature contributions to individual predictions (see Details).}
\item{approxcontrib}{whether to use a fast approximation for feature contributions (see Details).}
\item{approxcontrib}{Whether to use a fast approximation for feature contributions (see Details).}
\item{predinteraction}{whether to return contributions of feature interactions to individual predictions (see Details).}
\item{predinteraction}{Whether to return contributions of feature interactions to individual predictions (see Details).}
\item{reshape}{whether to reshape the vector of predictions to a matrix form when there are several
prediction outputs per case. This option has no effect when either of predleaf, predcontrib,
or predinteraction flags is TRUE.}
\item{reshape}{Whether to reshape the vector of predictions to matrix form when there are several
prediction outputs per case. No effect if \code{predleaf}, \code{predcontrib},
or \code{predinteraction} is \code{TRUE}.}
\item{training}{whether is the prediction result used for training. For dart booster,
\item{training}{Whether the predictions are used for training. For dart booster,
training predicting will perform dropout.}
\item{iterationrange}{Specifies which layer of trees are used in prediction. For
example, if a random forest is trained with 100 rounds. Specifying
`iterationrange=(1, 21)`, then only the forests built during [1, 21) (half open set)
rounds are used in this prediction. It's 1-based index just like R vector. When set
to \code{c(1, 1)} XGBoost will use all trees.}
\item{iterationrange}{Specifies which trees are used in prediction. For
example, take a random forest with 100 rounds.
With \code{iterationrange=c(1, 21)}, only the trees built during \verb{[1, 21)} (half open set)
rounds are used in this prediction. The index is 1-based just like an R vector. When set
to \code{c(1, 1)}, XGBoost will use all trees.}
\item{strict_shape}{Default is \code{FALSE}. When it's set to \code{TRUE}, output
type and shape of prediction are invariant to model type.}
\item{strict_shape}{Default is \code{FALSE}. When set to \code{TRUE}, the output
type and shape of predictions are invariant to the model type.}
\item{...}{Parameters passed to \code{predict.xgb.Booster}}
\item{...}{Not used.}
}
\value{
The return type is different depending whether \code{strict_shape} is set to \code{TRUE}. By default,
for regression or binary classification, it returns a vector of length \code{nrows(newdata)}.
For multiclass classification, either a \code{num_class * nrows(newdata)} vector or
a \code{(nrows(newdata), num_class)} dimension matrix is returned, depending on
the \code{reshape} value.
The return type depends on \code{strict_shape}. If \code{FALSE} (default):
\itemize{
\item For regression or binary classification: A vector of length \code{nrows(newdata)}.
\item For multiclass classification: A vector of length \code{num_class * nrows(newdata)} or
a \verb{(nrows(newdata), num_class)} matrix, depending on the \code{reshape} value.
\item When \code{predleaf = TRUE}: A matrix with one column per tree.
\item When \code{predcontrib = TRUE}: When not multiclass, a matrix with
\code{ num_features + 1} columns. The last "+ 1" column corresponds to the baseline value.
In the multiclass case, a list of \code{num_class} such matrices.
The contribution values are on the scale of untransformed margin
(e.g., for binary classification, the values are log-odds deviations from the baseline).
\item When \code{predinteraction = TRUE}: When not multiclass, the output is a 3d array of
dimension \code{c(nrow, num_features + 1, num_features + 1)}. The off-diagonal (in the last two dimensions)
elements represent different feature interaction contributions. The array is symmetric WRT the last
two dimensions. The "+ 1" columns corresponds to the baselines. Summing this array along the last dimension should
produce practically the same result as \code{predcontrib = TRUE}.
In the multiclass case, a list of \code{num_class} such arrays.
}
When \code{predleaf = TRUE}, the output is a matrix object with the
number of columns corresponding to the number of trees.
When \code{predcontrib = TRUE} and it is not a multiclass setting, the output is a matrix object with
\code{num_features + 1} columns. The last "+ 1" column in a matrix corresponds to bias.
For a multiclass case, a list of \code{num_class} elements is returned, where each element is
such a matrix. The contribution values are on the scale of untransformed margin
(e.g., for binary classification would mean that the contributions are log-odds deviations from bias).
When \code{predinteraction = TRUE} and it is not a multiclass setting, the output is a 3d array with
dimensions \code{c(nrow, num_features + 1, num_features + 1)}. The off-diagonal (in the last two dimensions)
elements represent different features interaction contributions. The array is symmetric WRT the last
two dimensions. The "+ 1" columns corresponds to bias. Summing this array along the last dimension should
produce practically the same result as predict with \code{predcontrib = TRUE}.
For a multiclass case, a list of \code{num_class} elements is returned, where each element is
such an array.
When \code{strict_shape} is set to \code{TRUE}, the output is always an array. For
normal prediction, the output is a 2-dimension array \code{(num_class, nrow(newdata))}.
For \code{predcontrib = TRUE}, output is \code{(ncol(newdata) + 1, num_class, nrow(newdata))}
For \code{predinteraction = TRUE}, output is \code{(ncol(newdata) + 1, ncol(newdata) + 1, num_class, nrow(newdata))}
For \code{predleaf = TRUE}, output is \code{(n_trees_in_forest, num_class, n_iterations, nrow(newdata))}
When \code{strict_shape = TRUE}, the output is always an array:
\itemize{
\item For normal predictions, the output has dimension \verb{(num_class, nrow(newdata))}.
\item For \code{predcontrib = TRUE}, the dimension is \verb{(ncol(newdata) + 1, num_class, nrow(newdata))}.
\item For \code{predinteraction = TRUE}, the dimension is \verb{(ncol(newdata) + 1, ncol(newdata) + 1, num_class, nrow(newdata))}.
\item For \code{predleaf = TRUE}, the dimension is \verb{(n_trees_in_forest, num_class, n_iterations, nrow(newdata))}.
}
}
\description{
Predicted values based on either xgboost model or model handle object.
}
\details{
Note that \code{iterationrange} would currently do nothing for predictions from gblinear,
since gblinear doesn't keep its boosting history.
Note that \code{iterationrange} would currently do nothing for predictions from "gblinear",
since "gblinear" doesn't keep its boosting history.
One possible practical applications of the \code{predleaf} option is to use the model
as a generator of new features which capture non-linearity and interactions,
e.g., as implemented in \code{\link{xgb.create.features}}.
e.g., as implemented in \code{\link[=xgb.create.features]{xgb.create.features()}}.
Setting \code{predcontrib = TRUE} allows to calculate contributions of each feature to
individual predictions. For "gblinear" booster, feature contributions are simply linear terms
@ -124,14 +117,14 @@ Since it quadratically depends on the number of features, it is recommended to p
of the most important features first. See below about the format of the returned results.
The \code{predict()} method uses as many threads as defined in \code{xgb.Booster} object (all by default).
If you want to change their number, then assign a new number to \code{nthread} using \code{\link{xgb.parameters<-}}.
Note also that converting a matrix to \code{\link{xgb.DMatrix}} uses multiple threads too.
If you want to change their number, assign a new number to \code{nthread} using \code{\link[=xgb.parameters<-]{xgb.parameters<-()}}.
Note that converting a matrix to \code{\link[=xgb.DMatrix]{xgb.DMatrix()}} uses multiple threads too.
}
\examples{
## binary classification:
data(agaricus.train, package='xgboost')
data(agaricus.test, package='xgboost')
data(agaricus.train, package = "xgboost")
data(agaricus.test, package = "xgboost")
## Keep the number of threads to 2 for examples
nthread <- 2
@ -140,8 +133,15 @@ data.table::setDTthreads(nthread)
train <- agaricus.train
test <- agaricus.test
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 0.5, nthread = nthread, nrounds = 5, objective = "binary:logistic")
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label),
max_depth = 2,
eta = 0.5,
nthread = nthread,
nrounds = 5,
objective = "binary:logistic"
)
# use all trees by default
pred <- predict(bst, test$data)
# use only the 1st tree
@ -173,39 +173,61 @@ par(mar = old_mar)
lb <- as.numeric(iris$Species) - 1
num_class <- 3
set.seed(11)
bst <- xgboost(data = as.matrix(iris[, -5]), label = lb,
max_depth = 4, eta = 0.5, nthread = 2, nrounds = 10, subsample = 0.5,
objective = "multi:softprob", num_class = num_class)
bst <- xgb.train(
data = xgb.DMatrix(as.matrix(iris[, -5]), label = lb),
max_depth = 4,
eta = 0.5,
nthread = 2,
nrounds = 10,
subsample = 0.5,
objective = "multi:softprob",
num_class = num_class
)
# predict for softmax returns num_class probability numbers per case:
pred <- predict(bst, as.matrix(iris[, -5]))
str(pred)
# reshape it to a num_class-columns matrix
pred <- matrix(pred, ncol=num_class, byrow=TRUE)
pred <- matrix(pred, ncol = num_class, byrow = TRUE)
# convert the probabilities to softmax labels
pred_labels <- max.col(pred) - 1
# the following should result in the same error as seen in the last iteration
sum(pred_labels != lb)/length(lb)
sum(pred_labels != lb) / length(lb)
# compare that to the predictions from softmax:
# compare with predictions from softmax:
set.seed(11)
bst <- xgboost(data = as.matrix(iris[, -5]), label = lb,
max_depth = 4, eta = 0.5, nthread = 2, nrounds = 10, subsample = 0.5,
objective = "multi:softmax", num_class = num_class)
bst <- xgb.train(
data = xgb.DMatrix(as.matrix(iris[, -5]), label = lb),
max_depth = 4,
eta = 0.5,
nthread = 2,
nrounds = 10,
subsample = 0.5,
objective = "multi:softmax",
num_class = num_class
)
pred <- predict(bst, as.matrix(iris[, -5]))
str(pred)
all.equal(pred, pred_labels)
# prediction from using only 5 iterations should result
# in the same error as seen in iteration 5:
pred5 <- predict(bst, as.matrix(iris[, -5]), iterationrange=c(1, 6))
sum(pred5 != lb)/length(lb)
pred5 <- predict(bst, as.matrix(iris[, -5]), iterationrange = c(1, 6))
sum(pred5 != lb) / length(lb)
}
\references{
Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions", NIPS Proceedings 2017, \url{https://arxiv.org/abs/1705.07874}
Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles", \url{https://arxiv.org/abs/1706.06060}
\enumerate{
\item Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions",
NIPS Proceedings 2017, \url{https://arxiv.org/abs/1705.07874}
\item Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles",
\url{https://arxiv.org/abs/1706.06060}
}
}
\seealso{
\code{\link{xgb.train}}.
\code{\link[=xgb.train]{xgb.train()}}
}

31
R-package/man/print.xgb.Booster.Rd
View File

@ -4,26 +4,35 @@
\alias{print.xgb.Booster}
\title{Print xgb.Booster}
\usage{
\method{print}{xgb.Booster}(x, verbose = FALSE, ...)
\method{print}{xgb.Booster}(x, ...)
}
\arguments{
\item{x}{an xgb.Booster object}
\item{x}{An \code{xgb.Booster} object.}
\item{verbose}{whether to print detailed data (e.g., attribute values)}
\item{...}{not currently used}
\item{...}{Not used.}
}
\value{
The same \code{x} object, returned invisibly
}
\description{
Print information about xgb.Booster.
Print information about \code{xgb.Booster}.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
train <- agaricus.train
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
attr(bst, 'myattr') <- 'memo'
bst <- xgboost(
data = train$data,
label = train$label,
max_depth = 2,
eta = 1,
nthread = 2,
nrounds = 2,
objective = "binary:logistic"
)
attr(bst, "myattr") <- "memo"
print(bst)
print(bst, verbose=TRUE)
}

42
R-package/man/setinfo.Rd
View File

@ -1,42 +0,0 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{setinfo}
\alias{setinfo}
\alias{setinfo.xgb.DMatrix}
\title{Set information of an xgb.DMatrix object}
\usage{
setinfo(object, ...)
\method{setinfo}{xgb.DMatrix}(object, name, info, ...)
}
\arguments{
\item{object}{Object of class "xgb.DMatrix"}
\item{...}{other parameters}
\item{name}{the name of the field to get}
\item{info}{the specific field of information to set}
}
\description{
Set information of an xgb.DMatrix object
}
\details{
The \code{name} field can be one of the following:
\itemize{
\item \code{label}: label XGBoost learn from ;
\item \code{weight}: to do a weight rescale ;
\item \code{base_margin}: base margin is the base prediction XGBoost will boost from ;
\item \code{group}: number of rows in each group (to use with \code{rank:pairwise} objective).
}
}
\examples{
data(agaricus.train, package='xgboost')
dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
labels <- getinfo(dtrain, 'label')
setinfo(dtrain, 'label', 1-labels)
labels2 <- getinfo(dtrain, 'label')
stopifnot(all.equal(labels2, 1-labels))
}

6
R-package/man/slice.xgb.DMatrix.Rd
View File

@ -7,17 +7,15 @@
\title{Get a new DMatrix containing the specified rows of
original xgb.DMatrix object}
\usage{
slice(object, ...)
slice(object, idxset)
\method{slice}{xgb.DMatrix}(object, idxset, ...)
\method{slice}{xgb.DMatrix}(object, idxset)
\method{[}{xgb.DMatrix}(object, idxset, colset = NULL)
}
\arguments{
\item{object}{Object of class "xgb.DMatrix"}
\item{...}{other parameters (currently not used)}
\item{idxset}{a integer vector of indices of rows needed}
\item{colset}{currently not used (columns subsetting is not available)}

22
R-package/man/variable.names.xgb.Booster.Rd Normal file
View File

@ -0,0 +1,22 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{variable.names.xgb.Booster}
\alias{variable.names.xgb.Booster}
\title{Get Features Names from Booster}
\usage{
\method{variable.names}{xgb.Booster}(object, ...)
}
\arguments{
\item{object}{An \code{xgb.Booster} object.}
\item{...}{Not used.}
}
\description{
Returns the feature / variable / column names from a fitted
booster object, which are set automatically during the call to \link{xgb.train}
from the DMatrix names, or which can be set manually through \link{setinfo}.
If the object doesn't have feature names, will return \code{NULL}.
It is equivalent to calling \code{getinfo(object, "feature_name")}.
}

52
R-package/man/xgb.Booster.complete.Rd
View File

@ -1,52 +0,0 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{xgb.Booster.complete}
\alias{xgb.Booster.complete}
\title{Restore missing parts of an incomplete xgb.Booster object.}
\usage{
xgb.Booster.complete(object, saveraw = TRUE)
}
\arguments{
\item{object}{object of class \code{xgb.Booster}}
\item{saveraw}{a flag indicating whether to append \code{raw} Booster memory dump data
when it doesn't already exist.}
}
\value{
An object of \code{xgb.Booster} class.
}
\description{
It attempts to complete an \code{xgb.Booster} object by restoring either its missing
raw model memory dump (when it has no \code{raw} data but its \code{xgb.Booster.handle} is valid)
or its missing internal handle (when its \code{xgb.Booster.handle} is not valid
but it has a raw Booster memory dump).
}
\details{
While this method is primarily for internal use, it might be useful in some practical situations.
E.g., when an \code{xgb.Booster} model is saved as an R object and then is loaded as an R object,
its handle (pointer) to an internal xgboost model would be invalid. The majority of xgboost methods
should still work for such a model object since those methods would be using
\code{xgb.Booster.complete} internally. However, one might find it to be more efficient to call the
\code{xgb.Booster.complete} function explicitly once after loading a model as an R-object.
That would prevent further repeated implicit reconstruction of an internal booster model.
}
\examples{
data(agaricus.train, package='xgboost')
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
saveRDS(bst, "xgb.model.rds")
# Warning: The resulting RDS file is only compatible with the current XGBoost version.
# Refer to the section titled "a-compatibility-note-for-saveRDS-save".
bst1 <- readRDS("xgb.model.rds")
if (file.exists("xgb.model.rds")) file.remove("xgb.model.rds")
# the handle is invalid:
print(bst1$handle)
bst1 <- xgb.Booster.complete(bst1)
# now the handle points to a valid internal booster model:
print(bst1$handle)
}

66
R-package/man/xgb.DMatrix.Rd
View File

@ -6,11 +6,19 @@
\usage{
xgb.DMatrix(
data,
info = list(),
label = NULL,
weight = NULL,
base_margin = NULL,
missing = NA,
silent = FALSE,
feature_names = colnames(data),
nthread = NULL,
...
group = NULL,
qid = NULL,
label_lower_bound = NULL,
label_upper_bound = NULL,
feature_weights = NULL,
enable_categorical = FALSE
)
}
\arguments{
@ -19,23 +27,65 @@ a \code{dgRMatrix} object,
a \code{dsparseVector} object (only when making predictions from a fitted model, will be
interpreted as a row vector), or a character string representing a filename.}
\item{info}{a named list of additional information to store in the \code{xgb.DMatrix} object.
See \code{\link{setinfo}} for the specific allowed kinds of}
\item{label}{Label of the training data.}
\item{weight}{Weight for each instance.
Note that, for ranking task, weights are per-group. In ranking task, one weight
is assigned to each group (not each data point). This is because we
only care about the relative ordering of data points within each group,
so it doesn't make sense to assign weights to individual data points.}
\item{base_margin}{Base margin used for boosting from existing model.
\if{html}{\out{<div class="sourceCode">}}\preformatted{ In the case of multi-output models, one can also pass multi-dimensional base_margin.
}\if{html}{\out{</div>}}}
\item{missing}{a float value to represents missing values in data (used only when input is a dense matrix).
It is useful when a 0 or some other extreme value represents missing values in data.}
\item{silent}{whether to suppress printing an informational message after loading from a file.}
\item{feature_names}{Set names for features. Overrides column names in data
frame and matrix.}
\item{nthread}{Number of threads used for creating DMatrix.}
\item{...}{the \code{info} data could be passed directly as parameters, without creating an \code{info} list.}
\item{group}{Group size for all ranking group.}
\item{qid}{Query ID for data samples, used for ranking.}
\item{label_lower_bound}{Lower bound for survival training.}
\item{label_upper_bound}{Upper bound for survival training.}
\item{feature_weights}{Set feature weights for column sampling.}
\item{enable_categorical}{Experimental support of specializing for categorical features.
\if{html}{\out{<div class="sourceCode">}}\preformatted{ If passing 'TRUE' and 'data' is a data frame,
columns of categorical types will automatically
be set to be of categorical type (feature_type='c') in the resulting DMatrix.
If passing 'FALSE' and 'data' is a data frame with categorical columns,
it will result in an error being thrown.
If 'data' is not a data frame, this argument is ignored.
JSON/UBJSON serialization format is required for this.
}\if{html}{\out{</div>}}}
}
\description{
Construct xgb.DMatrix object from either a dense matrix, a sparse matrix, or a local file.
Supported input file formats are either a LIBSVM text file or a binary file that was created previously by
\code{\link{xgb.DMatrix.save}}).
}
\details{
Note that DMatrix objects are not serializable through R functions such as \code{saveRDS} or \code{save}.
If a DMatrix gets serialized and then de-serialized (for example, when saving data in an R session or caching
chunks in an Rmd file), the resulting object will not be usable anymore and will need to be reconstructed
from the original source of data.
}
\examples{
data(agaricus.train, package='xgboost')
## Keep the number of threads to 1 for examples
@ -44,7 +94,7 @@ data.table::setDTthreads(nthread)
dtrain <- with(
agaricus.train, xgb.DMatrix(data, label = label, nthread = nthread)
)
xgb.DMatrix.save(dtrain, 'xgb.DMatrix.data')
dtrain <- xgb.DMatrix('xgb.DMatrix.data')
if (file.exists('xgb.DMatrix.data')) file.remove('xgb.DMatrix.data')
fname <- file.path(tempdir(), "xgb.DMatrix.data")
xgb.DMatrix.save(dtrain, fname)
dtrain <- xgb.DMatrix(fname)
}

32
R-package/man/xgb.DMatrix.hasinfo.Rd Normal file
View File

@ -0,0 +1,32 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{xgb.DMatrix.hasinfo}
\alias{xgb.DMatrix.hasinfo}
\title{Check whether DMatrix object has a field}
\usage{
xgb.DMatrix.hasinfo(object, info)
}
\arguments{
\item{object}{The DMatrix object to check for the given \code{info} field.}
\item{info}{The field to check for presence or absence in \code{object}.}
}
\description{
Checks whether an xgb.DMatrix object has a given field assigned to
it, such as weights, labels, etc.
}
\examples{
library(xgboost)
x <- matrix(1:10, nrow = 5)
dm <- xgb.DMatrix(x, nthread = 1)
# 'dm' so far doesn't have any fields set
xgb.DMatrix.hasinfo(dm, "label")
# Fields can be added after construction
setinfo(dm, "label", 1:5)
xgb.DMatrix.hasinfo(dm, "label")
}
\seealso{
\link{xgb.DMatrix}, \link{getinfo.xgb.DMatrix}, \link{setinfo.xgb.DMatrix}
}

6
R-package/man/xgb.DMatrix.save.Rd
View File

@ -17,7 +17,7 @@ Save xgb.DMatrix object to binary file
\examples{
data(agaricus.train, package='xgboost')
dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
xgb.DMatrix.save(dtrain, 'xgb.DMatrix.data')
dtrain <- xgb.DMatrix('xgb.DMatrix.data')
if (file.exists('xgb.DMatrix.data')) file.remove('xgb.DMatrix.data')
fname <- file.path(tempdir(), "xgb.DMatrix.data")
xgb.DMatrix.save(dtrain, fname)
dtrain <- xgb.DMatrix(fname)
}

61
R-package/man/xgb.attr.Rd
View File

@ -5,7 +5,7 @@
\alias{xgb.attr<-}
\alias{xgb.attributes}
\alias{xgb.attributes<-}
\title{Accessors for serializable attributes of a model.}
\title{Accessors for serializable attributes of a model}
\usage{
xgb.attr(object, name)
@ -16,64 +16,71 @@ xgb.attributes(object)
xgb.attributes(object) <- value
}
\arguments{
\item{object}{Object of class \code{xgb.Booster} or \code{xgb.Booster.handle}.}
\item{object}{Object of class \code{xgb.Booster}. \bold{Will be modified in-place} when assigning to it.}
\item{name}{a non-empty character string specifying which attribute is to be accessed.}
\item{name}{A non-empty character string specifying which attribute is to be accessed.}
\item{value}{a value of an attribute for \code{xgb.attr<-}; for \code{xgb.attributes<-}
it's a list (or an object coercible to a list) with the names of attributes to set
\item{value}{For \verb{xgb.attr<-}, a value of an attribute; for \verb{xgb.attributes<-},
it is a list (or an object coercible to a list) with the names of attributes to set
and the elements corresponding to attribute values.
Non-character values are converted to character.
When attribute value is not a scalar, only the first index is used.
When an attribute value is not a scalar, only the first index is used.
Use \code{NULL} to remove an attribute.}
}
\value{
\code{xgb.attr} returns either a string value of an attribute
\itemize{
\item \code{xgb.attr()} returns either a string value of an attribute
or \code{NULL} if an attribute wasn't stored in a model.
\code{xgb.attributes} returns a list of all attribute stored in a model
\item \code{xgb.attributes()} returns a list of all attributes stored in a model
or \code{NULL} if a model has no stored attributes.
}
}
\description{
These methods allow to manipulate the key-value attribute strings of an xgboost model.
}
\details{
The primary purpose of xgboost model attributes is to store some meta-data about the model.
The primary purpose of xgboost model attributes is to store some meta data about the model.
Note that they are a separate concept from the object attributes in R.
Specifically, they refer to key-value strings that can be attached to an xgboost model,
stored together with the model's binary representation, and accessed later
(from R or any other interface).
In contrast, any R-attribute assigned to an R-object of \code{xgb.Booster} class
would not be saved by \code{xgb.save} because an xgboost model is an external memory object
In contrast, any R attribute assigned to an R object of \code{xgb.Booster} class
would not be saved by \code{\link[=xgb.save]{xgb.save()}} because an xgboost model is an external memory object
and its serialization is handled externally.
Also, setting an attribute that has the same name as one of xgboost's parameters wouldn't
change the value of that parameter for a model.
Use \code{\link{xgb.parameters<-}} to set or change model parameters.
Use \code{\link[=xgb.parameters<-]{xgb.parameters<-()}} to set or change model parameters.
The attribute setters would usually work more efficiently for \code{xgb.Booster.handle}
than for \code{xgb.Booster}, since only just a handle (pointer) would need to be copied.
That would only matter if attributes need to be set many times.
Note, however, that when feeding a handle of an \code{xgb.Booster} object to the attribute setters,
the raw model cache of an \code{xgb.Booster} object would not be automatically updated,
and it would be user's responsibility to call \code{xgb.serialize} to update it.
The \code{xgb.attributes<-} setter either updates the existing or adds one or several attributes,
The \verb{xgb.attributes<-} setter either updates the existing or adds one or several attributes,
but it doesn't delete the other existing attributes.
Important: since this modifies the booster's C object, semantics for assignment here
will differ from R's, as any object reference to the same booster will be modified
too, while assignment of R attributes through \verb{attributes(model)$<attr> <- <value>}
will follow the usual copy-on-write R semantics (see \link{xgb.copy.Booster} for an
example of these behaviors).
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
train <- agaricus.train
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
bst <- xgboost(
data = train$data,
label = train$label,
max_depth = 2,
eta = 1,
nthread = 2,
nrounds = 2,
objective = "binary:logistic"
)
xgb.attr(bst, "my_attribute") <- "my attribute value"
print(xgb.attr(bst, "my_attribute"))
xgb.attributes(bst) <- list(a = 123, b = "abc")
xgb.save(bst, 'xgb.model')
bst1 <- xgb.load('xgb.model')
if (file.exists('xgb.model')) file.remove('xgb.model')
fname <- file.path(tempdir(), "xgb.ubj")
xgb.save(bst, fname)
bst1 <- xgb.load(fname)
print(xgb.attr(bst1, "my_attribute"))
print(xgb.attributes(bst1))

31
R-package/man/xgb.config.Rd
View File

@ -3,31 +3,48 @@
\name{xgb.config}
\alias{xgb.config}
\alias{xgb.config<-}
\title{Accessors for model parameters as JSON string.}
\title{Accessors for model parameters as JSON string}
\usage{
xgb.config(object)
xgb.config(object) <- value
}
\arguments{
\item{object}{Object of class \code{xgb.Booster}}
\item{object}{Object of class \code{xgb.Booster}. \bold{Will be modified in-place} when assigning to it.}
\item{value}{A JSON string.}
\item{value}{An R list.}
}
\value{
\code{xgb.config} will return the parameters as an R list.
}
\description{
Accessors for model parameters as JSON string.
Accessors for model parameters as JSON string
}
\details{
Note that assignment is performed in-place on the booster C object, which unlike assignment
of R attributes, doesn't follow typical copy-on-write semantics for assignment - i.e. all references
to the same booster will also get updated.
See \link{xgb.copy.Booster} for an example of this behavior.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
## Keep the number of threads to 1 for examples
nthread <- 1
data.table::setDTthreads(nthread)
train <- agaricus.train
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = nthread, nrounds = 2, objective = "binary:logistic"
data = train$data,
label = train$label,
max_depth = 2,
eta = 1,
nthread = nthread,
nrounds = 2,
objective = "binary:logistic"
)
config <- xgb.config(bst)
}

53
R-package/man/xgb.copy.Booster.Rd Normal file
View File

@ -0,0 +1,53 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{xgb.copy.Booster}
\alias{xgb.copy.Booster}
\title{Deep-copies a Booster Object}
\usage{
xgb.copy.Booster(model)
}
\arguments{
\item{model}{An 'xgb.Booster' object.}
}
\value{
A deep copy of \code{model} - it will be identical in every way, but C-level
functions called on that copy will not affect the \code{model} variable.
}
\description{
Creates a deep copy of an 'xgb.Booster' object, such that the
C object pointer contained will be a different object, and hence functions
like \link{xgb.attr} will not affect the object from which it was copied.
}
\examples{
library(xgboost)
data(mtcars)
y <- mtcars$mpg
x <- mtcars[, -1]
dm <- xgb.DMatrix(x, label = y, nthread = 1)
model <- xgb.train(
data = dm,
params = list(nthread = 1),
nround = 3
)
# Set an arbitrary attribute kept at the C level
xgb.attr(model, "my_attr") <- 100
print(xgb.attr(model, "my_attr"))
# Just assigning to a new variable will not create
# a deep copy - C object pointer is shared, and in-place
# modifications will affect both objects
model_shallow_copy <- model
xgb.attr(model_shallow_copy, "my_attr") <- 333
# 'model' was also affected by this change:
print(xgb.attr(model, "my_attr"))
model_deep_copy <- xgb.copy.Booster(model)
xgb.attr(model_deep_copy, "my_attr") <- 444
# 'model' was NOT affected by this change
# (keeps previous value that was assigned before)
print(xgb.attr(model, "my_attr"))
# Verify that the new object was actually modified
print(xgb.attr(model_deep_copy, "my_attr"))
}

4
R-package/man/xgb.create.features.Rd
View File

@ -48,7 +48,7 @@ be the binary vector \code{[0, 1, 0, 1, 0]}, where the first 3 entries
correspond to the leaves of the first subtree and last 2 to
those of the second subtree.
[...]
\link{...}
We can understand boosted decision tree
based transformation as a supervised feature encoding that
@ -62,7 +62,7 @@ data(agaricus.test, package='xgboost')
dtrain <- with(agaricus.train, xgb.DMatrix(data, label = label, nthread = 2))
dtest <- with(agaricus.test, xgb.DMatrix(data, label = label, nthread = 2))
param <- list(max_depth=2, eta=1, silent=1, objective='binary:logistic')
param <- list(max_depth=2, eta=1, objective='binary:logistic')
nrounds = 4
bst = xgb.train(params = param, data = dtrain, nrounds = nrounds, nthread = 2)

86
R-package/man/xgb.cv.Rd
View File

@ -29,22 +29,22 @@ xgb.cv(
}
\arguments{
\item{params}{the list of parameters. The complete list of parameters is
available in the \href{http://xgboost.readthedocs.io/en/latest/parameter.html}{online documentation}. Below
is a shorter summary:
available in the \href{http://xgboost.readthedocs.io/en/latest/parameter.html}{online documentation}. Below
is a shorter summary:
\itemize{
\item \code{objective} objective function, common ones are
\itemize{
\item \code{reg:squarederror} Regression with squared loss.
\item \code{binary:logistic} logistic regression for classification.
\item See \code{\link[=xgb.train]{xgb.train}()} for complete list of objectives.
}
\item \code{eta} step size of each boosting step
\item \code{max_depth} maximum depth of the tree
\item \code{nthread} number of thread used in training, if not set, all threads are used
\item \code{objective} objective function, common ones are
\itemize{
\item \code{reg:squarederror} Regression with squared loss.
\item \code{binary:logistic} logistic regression for classification.
\item See \code{\link[=xgb.train]{xgb.train}()} for complete list of objectives.
}
\item \code{eta} step size of each boosting step
\item \code{max_depth} maximum depth of the tree
\item \code{nthread} number of thread used in training, if not set, all threads are used
}
See \code{\link{xgb.train}} for further details.
See also demo/ for walkthrough example in R.}
See \code{\link{xgb.train}} for further details.
See also demo/ for walkthrough example in R.}
\item{data}{takes an \code{xgb.DMatrix}, \code{matrix}, or \code{dgCMatrix} as the input.}
@ -64,17 +64,17 @@ from each CV model. This parameter engages the \code{\link{cb.cv.predict}} callb
\item{showsd}{\code{boolean}, whether to show standard deviation of cross validation}
\item{metrics, }{list of evaluation metrics to be used in cross validation,
when it is not specified, the evaluation metric is chosen according to objective function.
Possible options are:
when it is not specified, the evaluation metric is chosen according to objective function.
Possible options are:
\itemize{
\item \code{error} binary classification error rate
\item \code{rmse} Rooted mean square error
\item \code{logloss} negative log-likelihood function
\item \code{mae} Mean absolute error
\item \code{mape} Mean absolute percentage error
\item \code{auc} Area under curve
\item \code{aucpr} Area under PR curve
\item \code{merror} Exact matching error, used to evaluate multi-class classification
\item \code{error} binary classification error rate
\item \code{rmse} Rooted mean square error
\item \code{logloss} negative log-likelihood function
\item \code{mae} Mean absolute error
\item \code{mape} Mean absolute percentage error
\item \code{auc} Area under curve
\item \code{aucpr} Area under PR curve
\item \code{merror} Exact matching error, used to evaluate multi-class classification
}}
\item{obj}{customized objective function. Returns gradient and second order
@ -120,26 +120,26 @@ to customize the training process.}
\value{
An object of class \code{xgb.cv.synchronous} with the following elements:
\itemize{
\item \code{call} a function call.
\item \code{params} parameters that were passed to the xgboost library. Note that it does not
capture parameters changed by the \code{\link{cb.reset.parameters}} callback.
\item \code{callbacks} callback functions that were either automatically assigned or
explicitly passed.
\item \code{evaluation_log} evaluation history stored as a \code{data.table} with the
first column corresponding to iteration number and the rest corresponding to the
CV-based evaluation means and standard deviations for the training and test CV-sets.
It is created by the \code{\link{cb.evaluation.log}} callback.
\item \code{niter} number of boosting iterations.
\item \code{nfeatures} number of features in training data.
\item \code{folds} the list of CV folds' indices - either those passed through the \code{folds}
parameter or randomly generated.
\item \code{best_iteration} iteration number with the best evaluation metric value
(only available with early stopping).
\item \code{best_ntreelimit} and the \code{ntreelimit} Deprecated attributes, use \code{best_iteration} instead.
\item \code{pred} CV prediction values available when \code{prediction} is set.
It is either vector or matrix (see \code{\link{cb.cv.predict}}).
\item \code{models} a list of the CV folds' models. It is only available with the explicit
setting of the \code{cb.cv.predict(save_models = TRUE)} callback.
\item \code{call} a function call.
\item \code{params} parameters that were passed to the xgboost library. Note that it does not
capture parameters changed by the \code{\link{cb.reset.parameters}} callback.
\item \code{callbacks} callback functions that were either automatically assigned or
explicitly passed.
\item \code{evaluation_log} evaluation history stored as a \code{data.table} with the
first column corresponding to iteration number and the rest corresponding to the
CV-based evaluation means and standard deviations for the training and test CV-sets.
It is created by the \code{\link{cb.evaluation.log}} callback.
\item \code{niter} number of boosting iterations.
\item \code{nfeatures} number of features in training data.
\item \code{folds} the list of CV folds' indices - either those passed through the \code{folds}
parameter or randomly generated.
\item \code{best_iteration} iteration number with the best evaluation metric value
(only available with early stopping).
\item \code{best_ntreelimit} and the \code{ntreelimit} Deprecated attributes, use \code{best_iteration} instead.
\item \code{pred} CV prediction values available when \code{prediction} is set.
It is either vector or matrix (see \code{\link{cb.cv.predict}}).
\item \code{models} a list of the CV folds' models. It is only available with the explicit
setting of the \code{cb.cv.predict(save_models = TRUE)} callback.
}
}
\description{

11
R-package/man/xgb.dump.Rd
View File

@ -9,7 +9,7 @@ xgb.dump(
fname = NULL,
fmap = "",
with_stats = FALSE,
dump_format = c("text", "json"),
dump_format = c("text", "json", "dot"),
...
)
}
@ -29,7 +29,10 @@ When this option is on, the model dump contains two additional values:
gain is the approximate loss function gain we get in each split;
cover is the sum of second order gradient in each node.}
\item{dump_format}{either 'text' or 'json' format could be specified.}
\item{dump_format}{either 'text', 'json', or 'dot' (graphviz) format could be specified.
Format 'dot' for a single tree can be passed directly to packages that consume this format
for graph visualization, such as function \code{\link[DiagrammeR:grViz]{DiagrammeR::grViz()}}}
\item{...}{currently not used}
}
@ -57,4 +60,8 @@ print(xgb.dump(bst, with_stats = TRUE))
# print in JSON format:
cat(xgb.dump(bst, with_stats = TRUE, dump_format='json'))
# plot first tree leveraging the 'dot' format
if (requireNamespace('DiagrammeR', quietly = TRUE)) {
DiagrammeR::grViz(xgb.dump(bst, dump_format = "dot")[[1L]])
}
}

11
R-package/man/xgb.gblinear.history.Rd
View File

@ -8,7 +8,8 @@ xgb.gblinear.history(model, class_index = NULL)
}
\arguments{
\item{model}{either an \code{xgb.Booster} or a result of \code{xgb.cv()}, trained
using the \code{cb.gblinear.history()} callback.}
using the \code{cb.gblinear.history()} callback, but \bold{not} a booster
loaded from \link{xgb.load} or \link{xgb.load.raw}.}
\item{class_index}{zero-based class index to extract the coefficients for only that
specific class in a multinomial multiclass model. When it is NULL, all the
@ -27,3 +28,11 @@ A helper function to extract the matrix of linear coefficients' history
from a gblinear model created while using the \code{cb.gblinear.history()}
callback.
}
\details{
Note that this is an R-specific function that relies on R attributes that
are not saved when using xgboost's own serialization functions like \link{xgb.load}
or \link{xgb.load.raw}.
In order for a serialized model to be accepted by tgis function, one must use R
serializers such as \link{saveRDS}.
}

19
R-package/man/xgb.get.DMatrix.data.Rd Normal file
View File

@ -0,0 +1,19 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{xgb.get.DMatrix.data}
\alias{xgb.get.DMatrix.data}
\title{Get DMatrix Data}
\usage{
xgb.get.DMatrix.data(dmat)
}
\arguments{
\item{dmat}{An \code{xgb.DMatrix} object, as returned by \link{xgb.DMatrix}.}
}
\value{
The data held in the DMatrix, as a sparse CSR matrix (class \code{dgRMatrix}
from package \code{Matrix}). If it had feature names, these will be added as column names
in the output.
}
\description{
Get DMatrix Data
}

17
R-package/man/xgb.get.DMatrix.num.non.missing.Rd Normal file
View File

@ -0,0 +1,17 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{xgb.get.DMatrix.num.non.missing}
\alias{xgb.get.DMatrix.num.non.missing}
\title{Get Number of Non-Missing Entries in DMatrix}
\usage{
xgb.get.DMatrix.num.non.missing(dmat)
}
\arguments{
\item{dmat}{An \code{xgb.DMatrix} object, as returned by \link{xgb.DMatrix}.}
}
\value{
The number of non-missing entries in the DMatrix
}
\description{
Get Number of Non-Missing Entries in DMatrix
}

58
R-package/man/xgb.get.DMatrix.qcut.Rd Normal file
View File

@ -0,0 +1,58 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.DMatrix.R
\name{xgb.get.DMatrix.qcut}
\alias{xgb.get.DMatrix.qcut}
\title{Get Quantile Cuts from DMatrix}
\usage{
xgb.get.DMatrix.qcut(dmat, output = c("list", "arrays"))
}
\arguments{
\item{dmat}{An \code{xgb.DMatrix} object, as returned by \link{xgb.DMatrix}.}
\item{output}{Output format for the quantile cuts. Possible options are:\itemize{
\item \code{"list"} will return the output as a list with one entry per column, where
each column will have a numeric vector with the cuts. The list will be named if
\code{dmat} has column names assigned to it.
\item \code{"arrays"} will return a list with entries \code{indptr} (base-0 indexing) and
\code{data}. Here, the cuts for column 'i' are obtained by slicing 'data' from entries
\code{indptr[i]+1} to \code{indptr[i+1]}.
}}
}
\value{
The quantile cuts, in the format specified by parameter \code{output}.
}
\description{
Get the quantile cuts (a.k.a. borders) from an \code{xgb.DMatrix}
that has been quantized for the histogram method (\code{tree_method="hist"}).
These cuts are used in order to assign observations to bins - i.e. these are ordered
boundaries which are used to determine assignment condition \verb{border_low < x < border_high}.
As such, the first and last bin will be outside of the range of the data, so as to include
all of the observations there.
If a given column has 'n' bins, then there will be 'n+1' cuts / borders for that column,
which will be output in sorted order from lowest to highest.
Different columns can have different numbers of bins according to their range.
}
\examples{
library(xgboost)
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(x, label = y, nthread = 1)
# DMatrix is not quantized right away, but will be once a hist model is generated
model <- xgb.train(
data = dm,
params = list(
tree_method = "hist",
max_bin = 8,
nthread = 1
),
nrounds = 3
)
# Now can get the quantile cuts
xgb.get.DMatrix.qcut(dm)
}

22
R-package/man/xgb.get.num.boosted.rounds.Rd Normal file
View File

@ -0,0 +1,22 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{xgb.get.num.boosted.rounds}
\alias{xgb.get.num.boosted.rounds}
\title{Get number of boosting in a fitted booster}
\usage{
xgb.get.num.boosted.rounds(model)
}
\arguments{
\item{model}{A fitted \code{xgb.Booster} model.}
}
\value{
The number of rounds saved in the model, as an integer.
}
\description{
Get number of boosting in a fitted booster
}
\details{
Note that setting booster parameters related to training
continuation / updates through \link{xgb.parameters<-} will reset the
number of rounds to zero.
}

134
R-package/man/xgb.importance.Rd
View File

@ -2,7 +2,7 @@
% Please edit documentation in R/xgb.importance.R
\name{xgb.importance}
\alias{xgb.importance}
\title{Importance of features in a model.}
\title{Feature importance}
\usage{
xgb.importance(
feature_names = NULL,
@ -14,88 +14,126 @@ xgb.importance(
)
}
\arguments{
\item{feature_names}{character vector of feature names. If the model already
contains feature names, those would be used when \code{feature_names=NULL} (default value).
Non-null \code{feature_names} could be provided to override those in the model.}
\item{feature_names}{Character vector used to overwrite the feature names
of the model. The default is \code{NULL} (use original feature names).}
\item{model}{object of class \code{xgb.Booster}.}
\item{model}{Object of class \code{xgb.Booster}.}
\item{trees}{(only for the gbtree booster) an integer vector of tree indices that should be included
into the importance calculation. If set to \code{NULL}, all trees of the model are parsed.
\item{trees}{An integer vector of tree indices that should be included
into the importance calculation (only for the "gbtree" booster).
The default (\code{NULL}) parses all trees.
It could be useful, e.g., in multiclass classification to get feature importances
for each class separately. IMPORTANT: the tree index in xgboost models
is zero-based (e.g., use \code{trees = 0:4} for first 5 trees).}
for each class separately. \emph{Important}: the tree index in XGBoost models
is zero-based (e.g., use \code{trees = 0:4} for the first five trees).}
\item{data}{deprecated.}
\item{data}{Deprecated.}
\item{label}{deprecated.}
\item{label}{Deprecated.}
\item{target}{deprecated.}
\item{target}{Deprecated.}
}
\value{
For a tree model, a \code{data.table} with the following columns:
A \code{data.table} with the following columns:
For a tree model:
\itemize{
\item \code{Features} names of the features used in the model;
\item \code{Gain} represents fractional contribution of each feature to the model based on
the total gain of this feature's splits. Higher percentage means a more important
predictive feature.
\item \code{Cover} metric of the number of observation related to this feature;
\item \code{Frequency} percentage representing the relative number of times
a feature have been used in trees.
\item \code{Features}: Names of the features used in the model.
\item \code{Gain}: Fractional contribution of each feature to the model based on
the total gain of this feature's splits. Higher percentage means higher importance.
\item \code{Cover}: Metric of the number of observation related to this feature.
\item \code{Frequency}: Percentage of times a feature has been used in trees.
}
A linear model's importance \code{data.table} has the following columns:
For a linear model:
\itemize{
\item \code{Features} names of the features used in the model;
\item \code{Weight} the linear coefficient of this feature;
\item \code{Class} (only for multiclass models) class label.
\item \code{Features}: Names of the features used in the model.
\item \code{Weight}: Linear coefficient of this feature.
\item \code{Class}: Class label (only for multiclass models).
}
If \code{feature_names} is not provided and \code{model} doesn't have \code{feature_names},
index of the features will be used instead. Because the index is extracted from the model dump
the index of the features will be used instead. Because the index is extracted from the model dump
(based on C++ code), it starts at 0 (as in C/C++ or Python) instead of 1 (usual in R).
}
\description{
Creates a \code{data.table} of feature importances in a model.
Creates a \code{data.table} of feature importances.
}
\details{
This function works for both linear and tree models.
For linear models, the importance is the absolute magnitude of linear coefficients.
For that reason, in order to obtain a meaningful ranking by importance for a linear model,
the features need to be on the same scale (which you also would want to do when using either
L1 or L2 regularization).
To obtain a meaningful ranking by importance for linear models, the features need to
be on the same scale (which is also recommended when using L1 or L2 regularization).
}
\examples{
# binomial classification using gbtree:
data(agaricus.train, package='xgboost')
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
# binomial classification using "gbtree":
data(agaricus.train, package = "xgboost")
bst <- xgboost(
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 2,
eta = 1,
nthread = 2,
nrounds = 2,
objective = "binary:logistic"
)
xgb.importance(model = bst)
# binomial classification using gblinear:
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, booster = "gblinear",
eta = 0.3, nthread = 1, nrounds = 20, objective = "binary:logistic")
# binomial classification using "gblinear":
bst <- xgboost(
data = agaricus.train$data,
label = agaricus.train$label,
booster = "gblinear",
eta = 0.3,
nthread = 1,
nrounds = 20,objective = "binary:logistic"
)
xgb.importance(model = bst)
# multiclass classification using gbtree:
# multiclass classification using "gbtree":
nclass <- 3
nrounds <- 10
mbst <- xgboost(data = as.matrix(iris[, -5]), label = as.numeric(iris$Species) - 1,
max_depth = 3, eta = 0.2, nthread = 2, nrounds = nrounds,
objective = "multi:softprob", num_class = nclass)
mbst <- xgboost(
data = as.matrix(iris[, -5]),
label = as.numeric(iris$Species) - 1,
max_depth = 3,
eta = 0.2,
nthread = 2,
nrounds = nrounds,
objective = "multi:softprob",
num_class = nclass
)
# all classes clumped together:
xgb.importance(model = mbst)
# inspect importances separately for each class:
xgb.importance(model = mbst, trees = seq(from=0, by=nclass, length.out=nrounds))
xgb.importance(model = mbst, trees = seq(from=1, by=nclass, length.out=nrounds))
xgb.importance(model = mbst, trees = seq(from=2, by=nclass, length.out=nrounds))
# multiclass classification using gblinear:
mbst <- xgboost(data = scale(as.matrix(iris[, -5])), label = as.numeric(iris$Species) - 1,
booster = "gblinear", eta = 0.2, nthread = 1, nrounds = 15,
objective = "multi:softprob", num_class = nclass)
# inspect importances separately for each class:
xgb.importance(
model = mbst, trees = seq(from = 0, by = nclass, length.out = nrounds)
)
xgb.importance(
model = mbst, trees = seq(from = 1, by = nclass, length.out = nrounds)
)
xgb.importance(
model = mbst, trees = seq(from = 2, by = nclass, length.out = nrounds)
)
# multiclass classification using "gblinear":
mbst <- xgboost(
data = scale(as.matrix(iris[, -5])),
label = as.numeric(iris$Species) - 1,
booster = "gblinear",
eta = 0.2,
nthread = 1,
nrounds = 15,
objective = "multi:softprob",
num_class = nclass
)
xgb.importance(model = mbst)
}

59
R-package/man/xgb.is.same.Booster.Rd Normal file
View File

@ -0,0 +1,59 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.Booster.R
\name{xgb.is.same.Booster}
\alias{xgb.is.same.Booster}
\title{Check if two boosters share the same C object}
\usage{
xgb.is.same.Booster(obj1, obj2)
}
\arguments{
\item{obj1}{Booster model to compare with \code{obj2}.}
\item{obj2}{Booster model to compare with \code{obj1}.}
}
\value{
Either \code{TRUE} or \code{FALSE} according to whether the two boosters share
the underlying C object.
}
\description{
Checks whether two booster objects refer to the same underlying C object.
}
\details{
As booster objects (as returned by e.g. \link{xgb.train}) contain an R 'externalptr'
object, they don't follow typical copy-on-write semantics of other R objects - that is, if
one assigns a booster to a different variable and modifies that new variable through in-place
methods like \link{xgb.attr<-}, the modification will be applied to both the old and the new
variable, unlike typical R assignments which would only modify the latter.
This function allows checking whether two booster objects share the same 'externalptr',
regardless of the R attributes that they might have.
In order to duplicate a booster in such a way that the copy wouldn't share the same
'externalptr', one can use function \link{xgb.copy.Booster}.
}
\examples{
library(xgboost)
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
model <- xgb.train(
params = list(nthread = 1),
data = xgb.DMatrix(x, label = y, nthread = 1),
nround = 3
)
model_shallow_copy <- model
xgb.is.same.Booster(model, model_shallow_copy) # same C object
model_deep_copy <- xgb.copy.Booster(model)
xgb.is.same.Booster(model, model_deep_copy) # different C objects
# In-place assignments modify all references,
# but not full/deep copies of the booster
xgb.attr(model_shallow_copy, "my_attr") <- 111
xgb.attr(model, "my_attr") # gets modified
xgb.attr(model_deep_copy, "my_attr") # doesn't get modified
}
\seealso{
\link{xgb.copy.Booster}
}

14
R-package/man/xgb.load.Rd
View File

@ -34,17 +34,19 @@ data.table::setDTthreads(nthread)
train <- agaricus.train
test <- agaricus.test
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2, eta = 1,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label),
max_depth = 2,
eta = 1,
nthread = nthread,
nrounds = 2,
objective = "binary:logistic"
)
xgb.save(bst, 'xgb.model')
bst <- xgb.load('xgb.model')
if (file.exists('xgb.model')) file.remove('xgb.model')
fname <- file.path(tempdir(), "xgb.ubj")
xgb.save(bst, fname)
bst <- xgb.load(fname)
}
\seealso{
\code{\link{xgb.save}}, \code{\link{xgb.Booster.complete}}.
\code{\link{xgb.save}}
}

4
R-package/man/xgb.load.raw.Rd
View File

@ -4,12 +4,10 @@
\alias{xgb.load.raw}
\title{Load serialised xgboost model from R's raw vector}
\usage{
xgb.load.raw(buffer, as_booster = FALSE)
xgb.load.raw(buffer)
}
\arguments{
\item{buffer}{the buffer returned by xgb.save.raw}
\item{as_booster}{Return the loaded model as xgb.Booster instead of xgb.Booster.handle.}
}
\description{
User can generate raw memory buffer by calling xgb.save.raw

90
R-package/man/xgb.model.dt.tree.Rd
View File

@ -2,7 +2,7 @@
% Please edit documentation in R/xgb.model.dt.tree.R
\name{xgb.model.dt.tree}
\alias{xgb.model.dt.tree}
\title{Parse a boosted tree model text dump}
\title{Parse model text dump}
\usage{
xgb.model.dt.tree(
feature_names = NULL,
@ -14,49 +14,48 @@ xgb.model.dt.tree(
)
}
\arguments{
\item{feature_names}{character vector of feature names. If the model already
contains feature names, those would be used when \code{feature_names=NULL} (default value).
Non-null \code{feature_names} could be provided to override those in the model.}
\item{feature_names}{Character vector of feature names. If the model already
contains feature names, those will be used when \code{feature_names=NULL} (default value).
\item{model}{object of class \code{xgb.Booster}}
\if{html}{\out{<div class="sourceCode">}}\preformatted{ Note that, if the model already contains feature names, it's \\bold\{not\} possible to override them here.
}\if{html}{\out{</div>}}}
\item{text}{\code{character} vector previously generated by the \code{xgb.dump}
function (where parameter \code{with_stats = TRUE} should have been set).
\code{text} takes precedence over \code{model}.}
\item{model}{Object of class \code{xgb.Booster}.}
\item{trees}{an integer vector of tree indices that should be parsed.
If set to \code{NULL}, all trees of the model are parsed.
It could be useful, e.g., in multiclass classification to get only
the trees of one certain class. IMPORTANT: the tree index in xgboost models
is zero-based (e.g., use \code{trees = 0:4} for first 5 trees).}
\item{text}{Character vector previously generated by the function \code{\link[=xgb.dump]{xgb.dump()}}
(called with parameter \code{with_stats = TRUE}). \code{text} takes precedence over \code{model}.}
\item{use_int_id}{a logical flag indicating whether nodes in columns "Yes", "No", "Missing" should be
represented as integers (when FALSE) or as "Tree-Node" character strings (when FALSE).}
\item{trees}{An integer vector of tree indices that should be used.
The default (\code{NULL}) uses all trees.
Useful, e.g., in multiclass classification to get only
the trees of one class. \emph{Important}: the tree index in XGBoost models
is zero-based (e.g., use \code{trees = 0:4} for the first five trees).}
\item{...}{currently not used.}
\item{use_int_id}{A logical flag indicating whether nodes in columns "Yes", "No", and
"Missing" should be represented as integers (when \code{TRUE}) or as "Tree-Node"
character strings (when \code{FALSE}, default).}
\item{...}{Currently not used.}
}
\value{
A \code{data.table} with detailed information about model trees' nodes.
The columns of the \code{data.table} are:
A \code{data.table} with detailed information about tree nodes. It has the following columns:
\itemize{
\item \code{Tree}: integer ID of a tree in a model (zero-based index)
\item \code{Node}: integer ID of a node in a tree (zero-based index)
\item \code{ID}: character identifier of a node in a model (only when \code{use_int_id=FALSE})
\item \code{Feature}: for a branch node, it's a feature id or name (when available);
for a leaf note, it simply labels it as \code{'Leaf'}
\item \code{Split}: location of the split for a branch node (split condition is always "less than")
\item \code{Yes}: ID of the next node when the split condition is met
\item \code{No}: ID of the next node when the split condition is not met
\item \code{Missing}: ID of the next node when branch value is missing
\item \code{Quality}: either the split gain (change in loss) or the leaf value
\item \code{Cover}: metric related to the number of observation either seen by a split
or collected by a leaf during training.
\item \code{Tree}: integer ID of a tree in a model (zero-based index).
\item \code{Node}: integer ID of a node in a tree (zero-based index).
\item \code{ID}: character identifier of a node in a model (only when \code{use_int_id = FALSE}).
\item \code{Feature}: for a branch node, a feature ID or name (when available);
for a leaf node, it simply labels it as \code{"Leaf"}.
\item \code{Split}: location of the split for a branch node (split condition is always "less than").
\item \code{Yes}: ID of the next node when the split condition is met.
\item \code{No}: ID of the next node when the split condition is not met.
\item \code{Missing}: ID of the next node when the branch value is missing.
\item \code{Gain}: either the split gain (change in loss) or the leaf value.
\item \code{Cover}: metric related to the number of observations either seen by a split
or collected by a leaf during training.
}
When \code{use_int_id=FALSE}, columns "Yes", "No", and "Missing" point to model-wide node identifiers
in the "ID" column. When \code{use_int_id=TRUE}, those columns point to node identifiers from
When \code{use_int_id = FALSE}, columns "Yes", "No", and "Missing" point to model-wide node identifiers
in the "ID" column. When \code{use_int_id = TRUE}, those columns point to node identifiers from
the corresponding trees in the "Node" column.
}
\description{
@ -65,22 +64,31 @@ Parse a boosted tree model text dump into a \code{data.table} structure.
\examples{
# Basic use:
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
## Keep the number of threads to 1 for examples
nthread <- 1
data.table::setDTthreads(nthread)
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 2,
eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
(dt <- xgb.model.dt.tree(colnames(agaricus.train$data), bst))
bst <- xgboost(
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 2,
eta = 1,
nthread = nthread,
nrounds = 2,
objective = "binary:logistic"
)
# This bst model already has feature_names stored with it, so those would be used when
# feature_names is not set:
(dt <- xgb.model.dt.tree(model = bst))
# How to match feature names of splits that are following a current 'Yes' branch:
merge(dt, dt[, .(ID, Y.Feature=Feature)], by.x='Yes', by.y='ID', all.x=TRUE)[order(Tree,Node)]
merge(
dt,
dt[, .(ID, Y.Feature = Feature)], by.x = "Yes", by.y = "ID", all.x = TRUE
)[
order(Tree, Node)
]
}

33
R-package/man/xgb.parameters.Rd
View File

@ -2,29 +2,46 @@
% Please edit documentation in R/xgb.Booster.R
\name{xgb.parameters<-}
\alias{xgb.parameters<-}
\title{Accessors for model parameters.}
\title{Accessors for model parameters}
\usage{
xgb.parameters(object) <- value
}
\arguments{
\item{object}{Object of class \code{xgb.Booster} or \code{xgb.Booster.handle}.}
\item{object}{Object of class \code{xgb.Booster}. \bold{Will be modified in-place}.}
\item{value}{a list (or an object coercible to a list) with the names of parameters to set
\item{value}{A list (or an object coercible to a list) with the names of parameters to set
and the elements corresponding to parameter values.}
}
\value{
The same booster \code{object}, which gets modified in-place.
}
\description{
Only the setter for xgboost parameters is currently implemented.
}
\details{
Note that the setter would usually work more efficiently for \code{xgb.Booster.handle}
than for \code{xgb.Booster}, since only just a handle would need to be copied.
Just like \link{xgb.attr}, this function will make in-place modifications
on the booster object which do not follow typical R assignment semantics - that is,
all references to the same booster will also be updated, unlike assingment of R
attributes which follow copy-on-write semantics.
See \link{xgb.copy.Booster} for an example of this behavior.
Be aware that setting parameters of a fitted booster related to training continuation / updates
will reset its number of rounds indicator to zero.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
train <- agaricus.train
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
bst <- xgboost(
data = train$data,
label = train$label,
max_depth = 2,
eta = 1,
nthread = 2,
nrounds = 2,
objective = "binary:logistic"
)
xgb.parameters(bst) <- list(eta = 0.1)

78
R-package/man/xgb.plot.deepness.Rd
View File

@ -3,7 +3,7 @@
\name{xgb.ggplot.deepness}
\alias{xgb.ggplot.deepness}
\alias{xgb.plot.deepness}
\title{Plot model trees deepness}
\title{Plot model tree depth}
\usage{
xgb.ggplot.deepness(
model = NULL,
@ -18,66 +18,84 @@ xgb.plot.deepness(
)
}
\arguments{
\item{model}{either an \code{xgb.Booster} model generated by the \code{xgb.train} function
or a data.table result of the \code{xgb.model.dt.tree} function.}
\item{model}{Either an \code{xgb.Booster} model, or the "data.table" returned by \code{\link[=xgb.model.dt.tree]{xgb.model.dt.tree()}}.}
\item{which}{which distribution to plot (see details).}
\item{which}{Which distribution to plot (see details).}
\item{plot}{(base R barplot) whether a barplot should be produced.
If FALSE, only a data.table is returned.}
\item{plot}{Should the plot be shown? Default is \code{TRUE}.}
\item{...}{other parameters passed to \code{barplot} or \code{plot}.}
\item{...}{Other parameters passed to \code{\link[graphics:barplot]{graphics::barplot()}} or \code{\link[graphics:plot.default]{graphics::plot()}}.}
}
\value{
Other than producing plots (when \code{plot=TRUE}), the \code{xgb.plot.deepness} function
silently returns a processed data.table where each row corresponds to a terminal leaf in a tree model,
and contains information about leaf's depth, cover, and weight (which is used in calculating predictions).
The \code{xgb.ggplot.deepness} silently returns either a list of two ggplot graphs when \code{which="2x1"}
or a single ggplot graph for the other \code{which} options.
The return value of the two functions is as follows:
\itemize{
\item \code{xgb.plot.deepness()}: A "data.table" (invisibly).
Each row corresponds to a terminal leaf in the model. It contains its information
about depth, cover, and weight (used in calculating predictions).
If \code{plot = TRUE}, also a plot is shown.
\item \code{xgb.ggplot.deepness()}: When \code{which = "2x1"}, a list of two "ggplot" objects,
and a single "ggplot" object otherwise.
}
}
\description{
Visualizes distributions related to depth of tree leafs.
\code{xgb.plot.deepness} uses base R graphics, while \code{xgb.ggplot.deepness} uses the ggplot backend.
Visualizes distributions related to the depth of tree leaves.
\itemize{
\item \code{xgb.plot.deepness()} uses base R graphics, while
\item \code{xgb.ggplot.deepness()} uses "ggplot2".
}
}
\details{
When \code{which="2x1"}, two distributions with respect to the leaf depth
When \code{which = "2x1"}, two distributions with respect to the leaf depth
are plotted on top of each other:
\itemize{
\item the distribution of the number of leafs in a tree model at a certain depth;
\item the distribution of average weighted number of observations ("cover")
ending up in leafs at certain depth.
\enumerate{
\item The distribution of the number of leaves in a tree model at a certain depth.
\item The distribution of the average weighted number of observations ("cover")
ending up in leaves at a certain depth.
}
Those could be helpful in determining sensible ranges of the \code{max_depth}
and \code{min_child_weight} parameters.
When \code{which="max.depth"} or \code{which="med.depth"}, plots of either maximum or median depth
per tree with respect to tree number are created. And \code{which="med.weight"} allows to see how
When \code{which = "max.depth"} or \code{which = "med.depth"}, plots of either maximum or
median depth per tree with respect to the tree number are created.
Finally, \code{which = "med.weight"} allows to see how
a tree's median absolute leaf weight changes through the iterations.
This function was inspired by the blog post
These functions have been inspired by the blog post
\url{https://github.com/aysent/random-forest-leaf-visualization}.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
## Keep the number of threads to 2 for examples
nthread <- 2
data.table::setDTthreads(nthread)
## Change max_depth to a higher number to get a more significant result
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 6,
eta = 0.1, nthread = nthread, nrounds = 50, objective = "binary:logistic",
subsample = 0.5, min_child_weight = 2)
bst <- xgboost(
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 6,
nthread = nthread,
nrounds = 50,
objective = "binary:logistic",
subsample = 0.5,
min_child_weight = 2
)
xgb.plot.deepness(bst)
xgb.ggplot.deepness(bst)
xgb.plot.deepness(bst, which='max.depth', pch=16, col=rgb(0,0,1,0.3), cex=2)
xgb.plot.deepness(
bst, which = "max.depth", pch = 16, col = rgb(0, 0, 1, 0.3), cex = 2
)
xgb.plot.deepness(bst, which='med.weight', pch=16, col=rgb(0,0,1,0.3), cex=2)
xgb.plot.deepness(
bst, which = "med.weight", pch = 16, col = rgb(0, 0, 1, 0.3), cex = 2
)
}
\seealso{
\code{\link{xgb.train}}, \code{\link{xgb.model.dt.tree}}.
\code{\link[=xgb.train]{xgb.train()}} and \code{\link[=xgb.model.dt.tree]{xgb.model.dt.tree()}}.
}

78
R-package/man/xgb.plot.importance.Rd
View File

@ -3,7 +3,7 @@
\name{xgb.ggplot.importance}
\alias{xgb.ggplot.importance}
\alias{xgb.plot.importance}
\title{Plot feature importance as a bar graph}
\title{Plot feature importance}
\usage{
xgb.ggplot.importance(
importance_matrix = NULL,
@ -26,74 +26,90 @@ xgb.plot.importance(
)
}
\arguments{
\item{importance_matrix}{a \code{data.table} returned by \code{\link{xgb.importance}}.}
\item{importance_matrix}{A \code{data.table} as returned by \code{\link[=xgb.importance]{xgb.importance()}}.}
\item{top_n}{maximal number of top features to include into the plot.}
\item{top_n}{Maximal number of top features to include into the plot.}
\item{measure}{the name of importance measure to plot.
\item{measure}{The name of importance measure to plot.
When \code{NULL}, 'Gain' would be used for trees and 'Weight' would be used for gblinear.}
\item{rel_to_first}{whether importance values should be represented as relative to the highest ranked feature.
See Details.}
\item{rel_to_first}{Whether importance values should be represented as relative to
the highest ranked feature, see Details.}
\item{n_clusters}{(ggplot only) a \code{numeric} vector containing the min and the max range
\item{n_clusters}{A numeric vector containing the min and the max range
of the possible number of clusters of bars.}
\item{...}{other parameters passed to \code{barplot} (except horiz, border, cex.names, names.arg, and las).}
\item{...}{Other parameters passed to \code{\link[graphics:barplot]{graphics::barplot()}}
(except \code{horiz}, \code{border}, \code{cex.names}, \code{names.arg}, and \code{las}).
Only used in \code{xgb.plot.importance()}.}
\item{left_margin}{(base R barplot) allows to adjust the left margin size to fit feature names.
When it is NULL, the existing \code{par('mar')} is used.}
\item{left_margin}{Adjust the left margin size to fit feature names.
When \code{NULL}, the existing \code{par("mar")} is used.}
\item{cex}{(base R barplot) passed as \code{cex.names} parameter to \code{barplot}.}
\item{cex}{Passed as \code{cex.names} parameter to \code{\link[graphics:barplot]{graphics::barplot()}}.}
\item{plot}{(base R barplot) whether a barplot should be produced.
If FALSE, only a data.table is returned.}
\item{plot}{Should the barplot be shown? Default is \code{TRUE}.}
}
\value{
The \code{xgb.plot.importance} function creates a \code{barplot} (when \code{plot=TRUE})
and silently returns a processed data.table with \code{n_top} features sorted by importance.
The \code{xgb.ggplot.importance} function returns a ggplot graph which could be customized afterwards.
E.g., to change the title of the graph, add \code{+ ggtitle("A GRAPH NAME")} to the result.
The return value depends on the function:
\itemize{
\item \code{xgb.plot.importance()}: Invisibly, a "data.table" with \code{n_top} features sorted
by importance. If \code{plot = TRUE}, the values are also plotted as barplot.
\item \code{xgb.ggplot.importance()}: A customizable "ggplot" object.
E.g., to change the title, set \code{+ ggtitle("A GRAPH NAME")}.
}
}
\description{
Represents previously calculated feature importance as a bar graph.
\code{xgb.plot.importance} uses base R graphics, while \code{xgb.ggplot.importance} uses the ggplot backend.
\itemize{
\item \code{xgb.plot.importance()} uses base R graphics, while
\item \code{xgb.ggplot.importance()} uses "ggplot".
}
}
\details{
The graph represents each feature as a horizontal bar of length proportional to the importance of a feature.
Features are shown ranked in a decreasing importance order.
It works for importances from both \code{gblinear} and \code{gbtree} models.
Features are sorted by decreasing importance.
It works for both "gblinear" and "gbtree" models.
When \code{rel_to_first = FALSE}, the values would be plotted as they were in \code{importance_matrix}.
For gbtree model, that would mean being normalized to the total of 1
When \code{rel_to_first = FALSE}, the values would be plotted as in \code{importance_matrix}.
For a "gbtree" model, that would mean being normalized to the total of 1
("what is feature's importance contribution relative to the whole model?").
For linear models, \code{rel_to_first = FALSE} would show actual values of the coefficients.
Setting \code{rel_to_first = TRUE} allows to see the picture from the perspective of
"what is feature's importance contribution relative to the most important feature?"
The ggplot-backend method also performs 1-D clustering of the importance values,
with bar colors corresponding to different clusters that have somewhat similar importance values.
The "ggplot" backend performs 1-D clustering of the importance values,
with bar colors corresponding to different clusters having similar importance values.
}
\examples{
data(agaricus.train)
## Keep the number of threads to 2 for examples
nthread <- 2
data.table::setDTthreads(nthread)
bst <- xgboost(
data = agaricus.train$data, label = agaricus.train$label, max_depth = 3,
eta = 1, nthread = nthread, nrounds = 2, objective = "binary:logistic"
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 3,
eta = 1,
nthread = nthread,
nrounds = 2,
objective = "binary:logistic"
)
importance_matrix <- xgb.importance(colnames(agaricus.train$data), model = bst)
xgb.plot.importance(
importance_matrix, rel_to_first = TRUE, xlab = "Relative importance"
)
xgb.plot.importance(importance_matrix, rel_to_first = TRUE, xlab = "Relative importance")
(gg <- xgb.ggplot.importance(importance_matrix, measure = "Frequency", rel_to_first = TRUE))
gg <- xgb.ggplot.importance(
importance_matrix, measure = "Frequency", rel_to_first = TRUE
)
gg
gg + ggplot2::ylab("Frequency")
}
\seealso{
\code{\link[graphics]{barplot}}.
\code{\link[graphics:barplot]{graphics::barplot()}}
}

58
R-package/man/xgb.plot.multi.trees.Rd
View File

@ -2,7 +2,7 @@
% Please edit documentation in R/xgb.plot.multi.trees.R
\name{xgb.plot.multi.trees}
\alias{xgb.plot.multi.trees}
\title{Project all trees on one tree and plot it}
\title{Project all trees on one tree}
\usage{
xgb.plot.multi.trees(
model,
@ -15,29 +15,31 @@ xgb.plot.multi.trees(
)
}
\arguments{
\item{model}{produced by the \code{xgb.train} function.}
\item{model}{Object of class \code{xgb.Booster}.}
\item{feature_names}{names of each feature as a \code{character} vector.}
\item{feature_names}{Character vector used to overwrite the feature names
of the model. The default (\code{NULL}) uses the original feature names.}
\item{features_keep}{number of features to keep in each position of the multi trees.}
\item{features_keep}{Number of features to keep in each position of the multi trees,
by default 5.}
\item{plot_width}{width in pixels of the graph to produce}
\item{plot_width, plot_height}{Width and height of the graph in pixels.
The values are passed to \code{\link[DiagrammeR:render_graph]{DiagrammeR::render_graph()}}.}
\item{plot_height}{height in pixels of the graph to produce}
\item{render}{Should the graph be rendered or not? The default is \code{TRUE}.}
\item{render}{a logical flag for whether the graph should be rendered (see Value).}
\item{...}{currently not used}
\item{...}{currently not used.}
}
\value{
When \code{render = TRUE}:
returns a rendered graph object which is an \code{htmlwidget} of class \code{grViz}.
Similar to ggplot objects, it needs to be printed to see it when not running from command line.
When \code{render = FALSE}:
silently returns a graph object which is of DiagrammeR's class \code{dgr_graph}.
The value depends on the \code{render} parameter:
\itemize{
\item If \code{render = TRUE} (default): Rendered graph object which is an htmlwidget of
class \code{grViz}. Similar to "ggplot" objects, it needs to be printed when not
running from the command line.
\item If \code{render = FALSE}: Graph object which is of DiagrammeR's class \code{dgr_graph}.
This could be useful if one wants to modify some of the graph attributes
before rendering the graph with \code{\link[DiagrammeR]{render_graph}}.
before rendering the graph with \code{\link[DiagrammeR:render_graph]{DiagrammeR::render_graph()}}.
}
}
\description{
Visualization of the ensemble of trees as a single collective unit.
@ -62,15 +64,22 @@ This function is inspired by this blog post:
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
## Keep the number of threads to 2 for examples
nthread <- 2
data.table::setDTthreads(nthread)
bst <- xgboost(
data = agaricus.train$data, label = agaricus.train$label, max_depth = 15,
eta = 1, nthread = nthread, nrounds = 30, objective = "binary:logistic",
min_child_weight = 50, verbose = 0
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 15,
eta = 1,
nthread = nthread,
nrounds = 30,
objective = "binary:logistic",
min_child_weight = 50,
verbose = 0
)
p <- xgb.plot.multi.trees(model = bst, features_keep = 3)
@ -78,10 +87,13 @@ print(p)
\dontrun{
# Below is an example of how to save this plot to a file.
# Note that for `export_graph` to work, the DiagrammeRsvg and rsvg packages must also be installed.
# Note that for export_graph() to work, the {DiagrammeRsvg} and {rsvg} packages
# must also be installed.
library(DiagrammeR)
gr <- xgb.plot.multi.trees(model=bst, features_keep = 3, render=FALSE)
export_graph(gr, 'tree.pdf', width=1500, height=600)
gr <- xgb.plot.multi.trees(model = bst, features_keep = 3, render = FALSE)
export_graph(gr, "tree.pdf", width = 1500, height = 600)
}
}

193
R-package/man/xgb.plot.shap.Rd
View File

@ -2,7 +2,7 @@
% Please edit documentation in R/xgb.plot.shap.R
\name{xgb.plot.shap}
\alias{xgb.plot.shap}
\title{SHAP contribution dependency plots}
\title{SHAP dependence plots}
\usage{
xgb.plot.shap(
data,
@ -33,87 +33,93 @@ xgb.plot.shap(
)
}
\arguments{
\item{data}{data as a \code{matrix} or \code{dgCMatrix}.}
\item{data}{The data to explain as a \code{matrix} or \code{dgCMatrix}.}
\item{shap_contrib}{a matrix of SHAP contributions that was computed earlier for the above
\code{data}. When it is NULL, it is computed internally using \code{model} and \code{data}.}
\item{shap_contrib}{Matrix of SHAP contributions of \code{data}.
The default (\code{NULL}) computes it from \code{model} and \code{data}.}
\item{features}{a vector of either column indices or of feature names to plot. When it is NULL,
feature importance is calculated, and \code{top_n} high ranked features are taken.}
\item{features}{Vector of column indices or feature names to plot.
When \code{NULL} (default), the \code{top_n} most important features are selected
by \code{\link[=xgb.importance]{xgb.importance()}}.}
\item{top_n}{when \code{features} is NULL, top_n [1, 100] most important features in a model are taken.}
\item{top_n}{How many of the most important features (<= 100) should be selected?
By default 1 for SHAP dependence and 10 for SHAP summary).
Only used when \code{features = NULL}.}
\item{model}{an \code{xgb.Booster} model. It has to be provided when either \code{shap_contrib}
or \code{features} is missing.}
\item{model}{An \code{xgb.Booster} model. Only required when \code{shap_contrib = NULL} or
\code{features = NULL}.}
\item{trees}{passed to \code{\link{xgb.importance}} when \code{features = NULL}.}
\item{trees}{Passed to \code{\link[=xgb.importance]{xgb.importance()}} when \code{features = NULL}.}
\item{target_class}{is only relevant for multiclass models. When it is set to a 0-based class index,
only SHAP contributions for that specific class are used.
If it is not set, SHAP importances are averaged over all classes.}
\item{target_class}{Only relevant for multiclass models. The default (\code{NULL})
averages the SHAP values over all classes. Pass a (0-based) class index
to show only SHAP values of that class.}
\item{approxcontrib}{passed to \code{\link{predict.xgb.Booster}} when \code{shap_contrib = NULL}.}
\item{approxcontrib}{Passed to \code{predict()} when \code{shap_contrib = NULL}.}
\item{subsample}{a random fraction of data points to use for plotting. When it is NULL,
it is set so that up to 100K data points are used.}
\item{subsample}{Fraction of data points randomly picked for plotting.
The default (\code{NULL}) will use up to 100k data points.}
\item{n_col}{a number of columns in a grid of plots.}
\item{n_col}{Number of columns in a grid of plots.}
\item{col}{color of the scatterplot markers.}
\item{col}{Color of the scatterplot markers.}
\item{pch}{scatterplot marker.}
\item{pch}{Scatterplot marker.}
\item{discrete_n_uniq}{a maximal number of unique values in a feature to consider it as discrete.}
\item{discrete_n_uniq}{Maximal number of unique feature values to consider the
feature as discrete.}
\item{discrete_jitter}{an \code{amount} parameter of jitter added to discrete features' positions.}
\item{discrete_jitter}{Jitter amount added to the values of discrete features.}
\item{ylab}{a y-axis label in 1D plots.}
\item{ylab}{The y-axis label in 1D plots.}
\item{plot_NA}{whether the contributions of cases with missing values should also be plotted.}
\item{plot_NA}{Should contributions of cases with missing values be plotted?
Default is \code{TRUE}.}
\item{col_NA}{a color of marker for missing value contributions.}
\item{col_NA}{Color of marker for missing value contributions.}
\item{pch_NA}{a marker type for NA values.}
\item{pch_NA}{Marker type for \code{NA} values.}
\item{pos_NA}{a relative position of the x-location where NA values are shown:
\item{pos_NA}{Relative position of the x-location where \code{NA} values are shown:
\code{min(x) + (max(x) - min(x)) * pos_NA}.}
\item{plot_loess}{whether to plot loess-smoothed curves. The smoothing is only done for features with
more than 5 distinct values.}
\item{plot_loess}{Should loess-smoothed curves be plotted? (Default is \code{TRUE}).
The smoothing is only done for features with more than 5 distinct values.}
\item{col_loess}{a color to use for the loess curves.}
\item{col_loess}{Color of loess curves.}
\item{span_loess}{the \code{span} parameter in \code{\link[stats]{loess}}'s call.}
\item{span_loess}{The \code{span} parameter of \code{\link[stats:loess]{stats::loess()}}.}
\item{which}{whether to do univariate or bivariate plotting. NOTE: only 1D is implemented so far.}
\item{which}{Whether to do univariate or bivariate plotting. Currently, only "1d" is implemented.}
\item{plot}{whether a plot should be drawn. If FALSE, only a list of matrices is returned.}
\item{plot}{Should the plot be drawn? (Default is \code{TRUE}).
If \code{FALSE}, only a list of matrices is returned.}
\item{...}{other parameters passed to \code{plot}.}
\item{...}{Other parameters passed to \code{\link[graphics:plot.default]{graphics::plot()}}.}
}
\value{
In addition to producing plots (when \code{plot=TRUE}), it silently returns a list of two matrices:
In addition to producing plots (when \code{plot = TRUE}), it silently returns a list of two matrices:
\itemize{
\item \code{data} the values of selected features;
\item \code{shap_contrib} the contributions of selected features.
\item \code{data}: Feature value matrix.
\item \code{shap_contrib}: Corresponding SHAP value matrix.
}
}
\description{
Visualizing the SHAP feature contribution to prediction dependencies on feature value.
Visualizes SHAP values against feature values to gain an impression of feature effects.
}
\details{
These scatterplots represent how SHAP feature contributions depend of feature values.
The similarity to partial dependency plots is that they also give an idea for how feature values
affect predictions. However, in partial dependency plots, we usually see marginal dependencies
of model prediction on feature value, while SHAP contribution dependency plots display the estimated
contributions of a feature to model prediction for each individual case.
The similarity to partial dependence plots is that they also give an idea for how feature values
affect predictions. However, in partial dependence plots, we see marginal dependencies
of model prediction on feature value, while SHAP dependence plots display the estimated
contributions of a feature to the prediction for each individual case.
When \code{plot_loess = TRUE} is set, feature values are rounded to 3 significant digits and
weighted LOESS is computed and plotted, where weights are the numbers of data points
When \code{plot_loess = TRUE}, feature values are rounded to three significant digits and
weighted LOESS is computed and plotted, where the weights are the numbers of data points
at each rounded value.
Note: SHAP contributions are shown on the scale of model margin. E.g., for a logistic binomial objective,
the margin is prediction before a sigmoidal transform into probability-like values.
Note: SHAP contributions are on the scale of the model margin.
E.g., for a logistic binomial objective, the margin is on log-odds scale.
Also, since SHAP stands for "SHapley Additive exPlanation" (model prediction = sum of SHAP
contributions for all features + bias), depending on the objective used, transforming SHAP
contributions for a feature from the marginal to the prediction space is not necessarily
@ -121,44 +127,99 @@ a meaningful thing to do.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.test, package='xgboost')
data(agaricus.train, package = "xgboost")
data(agaricus.test, package = "xgboost")
## Keep the number of threads to 1 for examples
nthread <- 1
data.table::setDTthreads(nthread)
nrounds <- 20
bst <- xgboost(agaricus.train$data, agaricus.train$label, nrounds = nrounds,
eta = 0.1, max_depth = 3, subsample = .5,
method = "hist", objective = "binary:logistic", nthread = nthread, verbose = 0)
bst <- xgboost(
agaricus.train$data,
agaricus.train$label,
nrounds = nrounds,
eta = 0.1,
max_depth = 3,
subsample = 0.5,
objective = "binary:logistic",
nthread = nthread,
verbose = 0
)
xgb.plot.shap(agaricus.test$data, model = bst, features = "odor=none")
contr <- predict(bst, agaricus.test$data, predcontrib = TRUE)
xgb.plot.shap(agaricus.test$data, contr, model = bst, top_n = 12, n_col = 3)
xgb.ggplot.shap.summary(agaricus.test$data, contr, model = bst, top_n = 12) # Summary plot
# multiclass example - plots for each class separately:
# Summary plot
xgb.ggplot.shap.summary(agaricus.test$data, contr, model = bst, top_n = 12)
# Multiclass example - plots for each class separately:
nclass <- 3
x <- as.matrix(iris[, -5])
set.seed(123)
is.na(x[sample(nrow(x) * 4, 30)]) <- TRUE # introduce some missing values
mbst <- xgboost(data = x, label = as.numeric(iris$Species) - 1, nrounds = nrounds,
max_depth = 2, eta = 0.3, subsample = .5, nthread = nthread,
objective = "multi:softprob", num_class = nclass, verbose = 0)
trees0 <- seq(from=0, by=nclass, length.out=nrounds)
mbst <- xgboost(
data = x,
label = as.numeric(iris$Species) - 1,
nrounds = nrounds,
max_depth = 2,
eta = 0.3,
subsample = 0.5,
nthread = nthread,
objective = "multi:softprob",
num_class = nclass,
verbose = 0
)
trees0 <- seq(from = 0, by = nclass, length.out = nrounds)
col <- rgb(0, 0, 1, 0.5)
xgb.plot.shap(x, model = mbst, trees = trees0, target_class = 0, top_n = 4,
n_col = 2, col = col, pch = 16, pch_NA = 17)
xgb.plot.shap(x, model = mbst, trees = trees0 + 1, target_class = 1, top_n = 4,
n_col = 2, col = col, pch = 16, pch_NA = 17)
xgb.plot.shap(x, model = mbst, trees = trees0 + 2, target_class = 2, top_n = 4,
n_col = 2, col = col, pch = 16, pch_NA = 17)
xgb.ggplot.shap.summary(x, model = mbst, target_class = 0, top_n = 4) # Summary plot
xgb.plot.shap(
x,
model = mbst,
trees = trees0,
target_class = 0,
top_n = 4,
n_col = 2,
col = col,
pch = 16,
pch_NA = 17
)
xgb.plot.shap(
x,
model = mbst,
trees = trees0 + 1,
target_class = 1,
top_n = 4,
n_col = 2,
col = col,
pch = 16,
pch_NA = 17
)
xgb.plot.shap(
x,
model = mbst,
trees = trees0 + 2,
target_class = 2,
top_n = 4,
n_col = 2,
col = col,
pch = 16,
pch_NA = 17
)
# Summary plot
xgb.ggplot.shap.summary(x, model = mbst, target_class = 0, top_n = 4)
}
\references{
Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions", NIPS Proceedings 2017, \url{https://arxiv.org/abs/1705.07874}
Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles", \url{https://arxiv.org/abs/1706.06060}
\enumerate{
\item Scott M. Lundberg, Su-In Lee, "A Unified Approach to Interpreting Model Predictions",
NIPS Proceedings 2017, \url{https://arxiv.org/abs/1705.07874}
\item Scott M. Lundberg, Su-In Lee, "Consistent feature attribution for tree ensembles",
\url{https://arxiv.org/abs/1706.06060}
}
}

55
R-package/man/xgb.plot.shap.summary.Rd
View File

@ -3,7 +3,7 @@
\name{xgb.ggplot.shap.summary}
\alias{xgb.ggplot.shap.summary}
\alias{xgb.plot.shap.summary}
\title{SHAP contribution dependency summary plot}
\title{SHAP summary plot}
\usage{
xgb.ggplot.shap.summary(
data,
@ -30,49 +30,54 @@ xgb.plot.shap.summary(
)
}
\arguments{
\item{data}{data as a \code{matrix} or \code{dgCMatrix}.}
\item{data}{The data to explain as a \code{matrix} or \code{dgCMatrix}.}
\item{shap_contrib}{a matrix of SHAP contributions that was computed earlier for the above
\code{data}. When it is NULL, it is computed internally using \code{model} and \code{data}.}
\item{shap_contrib}{Matrix of SHAP contributions of \code{data}.
The default (\code{NULL}) computes it from \code{model} and \code{data}.}
\item{features}{a vector of either column indices or of feature names to plot. When it is NULL,
feature importance is calculated, and \code{top_n} high ranked features are taken.}
\item{features}{Vector of column indices or feature names to plot.
When \code{NULL} (default), the \code{top_n} most important features are selected
by \code{\link[=xgb.importance]{xgb.importance()}}.}
\item{top_n}{when \code{features} is NULL, top_n [1, 100] most important features in a model are taken.}
\item{top_n}{How many of the most important features (<= 100) should be selected?
By default 1 for SHAP dependence and 10 for SHAP summary).
Only used when \code{features = NULL}.}
\item{model}{an \code{xgb.Booster} model. It has to be provided when either \code{shap_contrib}
or \code{features} is missing.}
\item{model}{An \code{xgb.Booster} model. Only required when \code{shap_contrib = NULL} or
\code{features = NULL}.}
\item{trees}{passed to \code{\link{xgb.importance}} when \code{features = NULL}.}
\item{trees}{Passed to \code{\link[=xgb.importance]{xgb.importance()}} when \code{features = NULL}.}
\item{target_class}{is only relevant for multiclass models. When it is set to a 0-based class index,
only SHAP contributions for that specific class are used.
If it is not set, SHAP importances are averaged over all classes.}
\item{target_class}{Only relevant for multiclass models. The default (\code{NULL})
averages the SHAP values over all classes. Pass a (0-based) class index
to show only SHAP values of that class.}
\item{approxcontrib}{passed to \code{\link{predict.xgb.Booster}} when \code{shap_contrib = NULL}.}
\item{approxcontrib}{Passed to \code{predict()} when \code{shap_contrib = NULL}.}
\item{subsample}{a random fraction of data points to use for plotting. When it is NULL,
it is set so that up to 100K data points are used.}
\item{subsample}{Fraction of data points randomly picked for plotting.
The default (\code{NULL}) will use up to 100k data points.}
}
\value{
A \code{ggplot2} object.
}
\description{
Compare SHAP contributions of different features.
Visualizes SHAP contributions of different features.
}
\details{
A point plot (each point representing one sample from \code{data}) is
A point plot (each point representing one observation from \code{data}) is
produced for each feature, with the points plotted on the SHAP value axis.
Each point (observation) is coloured based on its feature value. The plot
hence allows us to see which features have a negative / positive contribution
Each point (observation) is coloured based on its feature value.
The plot allows to see which features have a negative / positive contribution
on the model prediction, and whether the contribution is different for larger
or smaller values of the feature. We effectively try to replicate the
\code{summary_plot} function from https://github.com/shap/shap.
or smaller values of the feature. Inspired by the summary plot of
\url{https://github.com/shap/shap}.
}
\examples{
# See \code{\link{xgb.plot.shap}}.
# See examples in xgb.plot.shap()
}
\seealso{
\code{\link{xgb.plot.shap}}, \code{\link{xgb.ggplot.shap.summary}},
\url{https://github.com/shap/shap}
\code{\link[=xgb.plot.shap]{xgb.plot.shap()}}, \code{\link[=xgb.ggplot.shap.summary]{xgb.ggplot.shap.summary()}},
and the Python library \url{https://github.com/shap/shap}.
}

122
R-package/man/xgb.plot.tree.Rd
View File

@ -2,7 +2,7 @@
% Please edit documentation in R/xgb.plot.tree.R
\name{xgb.plot.tree}
\alias{xgb.plot.tree}
\title{Plot a boosted tree model}
\title{Plot boosted trees}
\usage{
xgb.plot.tree(
feature_names = NULL,
@ -12,80 +12,124 @@ xgb.plot.tree(
plot_height = NULL,
render = TRUE,
show_node_id = FALSE,
style = c("R", "xgboost"),
...
)
}
\arguments{
\item{feature_names}{names of each feature as a \code{character} vector.}
\item{feature_names}{Character vector used to overwrite the feature names
of the model. The default (\code{NULL}) uses the original feature names.}
\item{model}{produced by the \code{xgb.train} function.}
\item{model}{Object of class \code{xgb.Booster}.}
\item{trees}{an integer vector of tree indices that should be visualized.
If set to \code{NULL}, all trees of the model are included.
IMPORTANT: the tree index in xgboost model is zero-based
(e.g., use \code{trees = 0:2} for the first 3 trees in a model).}
\item{trees}{An integer vector of tree indices that should be used.
The default (\code{NULL}) uses all trees.
Useful, e.g., in multiclass classification to get only
the trees of one class. \emph{Important}: the tree index in XGBoost models
is zero-based (e.g., use \code{trees = 0:2} for the first three trees).}
\item{plot_width}{the width of the diagram in pixels.}
\item{plot_width, plot_height}{Width and height of the graph in pixels.
The values are passed to \code{\link[DiagrammeR:render_graph]{DiagrammeR::render_graph()}}.}
\item{plot_height}{the height of the diagram in pixels.}
\item{render}{a logical flag for whether the graph should be rendered (see Value).}
\item{render}{Should the graph be rendered or not? The default is \code{TRUE}.}
\item{show_node_id}{a logical flag for whether to show node id's in the graph.}
\item{style}{Style to use for the plot. Options are:\itemize{
\item \code{"xgboost"}: will use the plot style defined in the core XGBoost library,
which is shared between different interfaces through the 'dot' format. This
style was not available before version 2.1.0 in R. It always plots the trees
vertically (from top to bottom).
\item \code{"R"}: will use the style defined from XGBoost's R interface, which predates
the introducition of the standardized style from the core library. It might plot
the trees horizontally (from left to right).
}
Note that \code{style="xgboost"} is only supported when all of the following conditions are met:\itemize{
\item Only a single tree is being plotted.
\item Node IDs are not added to the graph.
\item The graph is being returned as \code{htmlwidget} (\code{render=TRUE}).
}}
\item{...}{currently not used.}
}
\value{
When \code{render = TRUE}:
returns a rendered graph object which is an \code{htmlwidget} of class \code{grViz}.
Similar to ggplot objects, it needs to be printed to see it when not running from command line.
When \code{render = FALSE}:
silently returns a graph object which is of DiagrammeR's class \code{dgr_graph}.
The value depends on the \code{render} parameter:
\itemize{
\item If \code{render = TRUE} (default): Rendered graph object which is an htmlwidget of
class \code{grViz}. Similar to "ggplot" objects, it needs to be printed when not
running from the command line.
\item If \code{render = FALSE}: Graph object which is of DiagrammeR's class \code{dgr_graph}.
This could be useful if one wants to modify some of the graph attributes
before rendering the graph with \code{\link[DiagrammeR]{render_graph}}.
before rendering the graph with \code{\link[DiagrammeR:render_graph]{DiagrammeR::render_graph()}}.
}
}
\description{
Read a tree model text dump and plot the model.
}
\details{
The content of each node is organised that way:
When using \code{style="xgboost"}, the content of each node is visualized as follows:
\itemize{
\item Feature name.
\item \code{Cover}: The sum of second order gradient of training data classified to the leaf.
If it is square loss, this simply corresponds to the number of instances seen by a split
or collected by a leaf during training.
The deeper in the tree a node is, the lower this metric will be.
\item \code{Gain} (for split nodes): the information gain metric of a split
(corresponds to the importance of the node in the model).
\item \code{Value} (for leafs): the margin value that the leaf may contribute to prediction.
\item For non-terminal nodes, it will display the split condition (number or name if
available, and the condition that would decide to which node to go next).
\item Those nodes will be connected to their children by arrows that indicate whether the
branch corresponds to the condition being met or not being met.
\item Terminal (leaf) nodes contain the margin to add when ending there.
}
The tree root nodes also indicate the Tree index (0-based).
When using \code{style="R"}, the content of each node is visualized like this:
\itemize{
\item \emph{Feature name}.
\item \emph{Cover:} The sum of second order gradients of training data.
For the squared loss, this simply corresponds to the number of instances in the node.
The deeper in the tree, the lower the value.
\item \emph{Gain} (for split nodes): Information gain metric of a split
(corresponds to the importance of the node in the model).
\item \emph{Value} (for leaves): Margin value that the leaf may contribute to the prediction.
}
The tree root nodes also indicate the tree index (0-based).
The "Yes" branches are marked by the "< split_value" label.
The branches that also used for missing values are marked as bold
The branches also used for missing values are marked as bold
(as in "carrying extra capacity").
This function uses \href{https://www.graphviz.org/}{GraphViz} as a backend of DiagrammeR.
This function uses \href{https://www.graphviz.org/}{GraphViz} as DiagrammeR backend.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.train, package = "xgboost")
bst <- xgboost(
data = agaricus.train$data,
label = agaricus.train$label,
max_depth = 3,
eta = 1,
nthread = 2,
nrounds = 2,
objective = "binary:logistic"
)
# plot the first tree, using the style from xgboost's core library
# (this plot should look identical to the ones generated from other
# interfaces like the python package for xgboost)
xgb.plot.tree(model = bst, trees = 1, style = "xgboost")
bst <- xgboost(data = agaricus.train$data, label = agaricus.train$label, max_depth = 3,
eta = 1, nthread = 2, nrounds = 2,objective = "binary:logistic")
# plot all the trees
xgb.plot.tree(model = bst)
xgb.plot.tree(model = bst, trees = NULL)
# plot only the first tree and display the node ID:
xgb.plot.tree(model = bst, trees = 0, show_node_id = TRUE)
\dontrun{
# Below is an example of how to save this plot to a file.
# Note that for `export_graph` to work, the DiagrammeRsvg and rsvg packages must also be installed.
# Note that for export_graph() to work, the {DiagrammeRsvg}
# and {rsvg} packages must also be installed.
library(DiagrammeR)
gr <- xgb.plot.tree(model=bst, trees=0:1, render=FALSE)
export_graph(gr, 'tree.pdf', width=1500, height=1900)
export_graph(gr, 'tree.png', width=1500, height=1900)
gr <- xgb.plot.tree(model = bst, trees = 0:1, render = FALSE)
export_graph(gr, "tree.pdf", width = 1500, height = 1900)
export_graph(gr, "tree.png", width = 1500, height = 1900)
}
}

36
R-package/man/xgb.save.Rd
View File

@ -7,15 +7,27 @@
xgb.save(model, fname)
}
\arguments{
\item{model}{model object of \code{xgb.Booster} class.}
\item{model}{Model object of \code{xgb.Booster} class.}
\item{fname}{name of the file to write.}
\item{fname}{Name of the file to write.
Note that the extension of this file name determined the serialization format to use:\itemize{
\item Extension ".ubj" will use the universal binary JSON format (recommended).
This format uses binary types for e.g. floating point numbers, thereby preventing any loss
of precision when converting to a human-readable JSON text or similar.
\item Extension ".json" will use plain JSON, which is a human-readable format.
\item Extension ".deprecated" will use a \bold{deprecated} binary format. This format will
not be able to save attributes introduced after v1 of XGBoost, such as the "best_iteration"
attribute that boosters might keep, nor feature names or user-specifiec attributes.
\item If the format is not specified by passing one of the file extensions above, will
default to UBJ.
}}
}
\description{
Save xgboost model to a file in binary format.
Save xgboost model to a file in binary or JSON format.
}
\details{
This methods allows to save a model in an xgboost-internal binary format which is universal
This methods allows to save a model in an xgboost-internal binary or text format which is universal
among the various xgboost interfaces. In R, the saved model file could be read-in later
using either the \code{\link{xgb.load}} function or the \code{xgb_model} parameter
of \code{\link{xgb.train}}.
@ -23,7 +35,7 @@ of \code{\link{xgb.train}}.
Note: a model can also be saved as an R-object (e.g., by using \code{\link[base]{readRDS}}
or \code{\link[base]{save}}). However, it would then only be compatible with R, and
corresponding R-methods would need to be used to load it. Moreover, persisting the model with
\code{\link[base]{readRDS}} or \code{\link[base]{save}}) will cause compatibility problems in
\code{\link[base]{readRDS}} or \code{\link[base]{save}}) might cause compatibility problems in
future versions of XGBoost. Consult \code{\link{a-compatibility-note-for-saveRDS-save}} to learn
how to persist models in a future-proof way, i.e. to make the model accessible in future
releases of XGBoost.
@ -38,16 +50,18 @@ data.table::setDTthreads(nthread)
train <- agaricus.train
test <- agaricus.test
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2, eta = 1,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label),
max_depth = 2,
eta = 1,
nthread = nthread,
nrounds = 2,
objective = "binary:logistic"
)
xgb.save(bst, 'xgb.model')
bst <- xgb.load('xgb.model')
if (file.exists('xgb.model')) file.remove('xgb.model')
fname <- file.path(tempdir(), "xgb.ubj")
xgb.save(bst, fname)
bst <- xgb.load(fname)
}
\seealso{
\code{\link{xgb.load}}, \code{\link{xgb.Booster.complete}}.
\code{\link{xgb.load}}
}

16
R-package/man/xgb.save.raw.Rd
View File

@ -5,19 +5,17 @@
\title{Save xgboost model to R's raw vector,
user can call xgb.load.raw to load the model back from raw vector}
\usage{
xgb.save.raw(model, raw_format = "deprecated")
xgb.save.raw(model, raw_format = "ubj")
}
\arguments{
\item{model}{the model object.}
\item{raw_format}{The format for encoding the booster. Available options are
\itemize{
\item \code{json}: Encode the booster into JSON text document.
\item \code{ubj}: Encode the booster into Universal Binary JSON.
\item \code{deprecated}: Encode the booster into old customized binary format.
}
Right now the default is \code{deprecated} but will be changed to \code{ubj} in upcoming release.}
\item \code{json}: Encode the booster into JSON text document.
\item \code{ubj}: Encode the booster into Universal Binary JSON.
\item \code{deprecated}: Encode the booster into old customized binary format.
}}
}
\description{
Save xgboost model from xgboost or xgb.train
@ -32,8 +30,8 @@ data.table::setDTthreads(nthread)
train <- agaricus.train
test <- agaricus.test
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
bst <- xgb.train(data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
eta = 1, nthread = nthread, nrounds = 2,objective = "binary:logistic")
raw <- xgb.save.raw(bst)
bst <- xgb.load.raw(raw)

29
R-package/man/xgb.serialize.Rd
View File

@ -1,29 +0,0 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.serialize.R
\name{xgb.serialize}
\alias{xgb.serialize}
\title{Serialize the booster instance into R's raw vector. The serialization method differs
from \code{\link{xgb.save.raw}} as the latter one saves only the model but not
parameters. This serialization format is not stable across different xgboost versions.}
\usage{
xgb.serialize(booster)
}
\arguments{
\item{booster}{the booster instance}
}
\description{
Serialize the booster instance into R's raw vector. The serialization method differs
from \code{\link{xgb.save.raw}} as the latter one saves only the model but not
parameters. This serialization format is not stable across different xgboost versions.
}
\examples{
data(agaricus.train, package='xgboost')
data(agaricus.test, package='xgboost')
train <- agaricus.train
test <- agaricus.test
bst <- xgboost(data = train$data, label = train$label, max_depth = 2,
eta = 1, nthread = 2, nrounds = 2,objective = "binary:logistic")
raw <- xgb.serialize(bst)
bst <- xgb.unserialize(raw)
}

55
R-package/man/xgb.shap.data.Rd
View File

@ -1,55 +0,0 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.plot.shap.R
\name{xgb.shap.data}
\alias{xgb.shap.data}
\title{Prepare data for SHAP plots. To be used in xgb.plot.shap, xgb.plot.shap.summary, etc.
Internal utility function.}
\usage{
xgb.shap.data(
data,
shap_contrib = NULL,
features = NULL,
top_n = 1,
model = NULL,
trees = NULL,
target_class = NULL,
approxcontrib = FALSE,
subsample = NULL,
max_observations = 1e+05
)
}
\arguments{
\item{data}{data as a \code{matrix} or \code{dgCMatrix}.}
\item{shap_contrib}{a matrix of SHAP contributions that was computed earlier for the above
\code{data}. When it is NULL, it is computed internally using \code{model} and \code{data}.}
\item{features}{a vector of either column indices or of feature names to plot. When it is NULL,
feature importance is calculated, and \code{top_n} high ranked features are taken.}
\item{top_n}{when \code{features} is NULL, top_n [1, 100] most important features in a model are taken.}
\item{model}{an \code{xgb.Booster} model. It has to be provided when either \code{shap_contrib}
or \code{features} is missing.}
\item{trees}{passed to \code{\link{xgb.importance}} when \code{features = NULL}.}
\item{target_class}{is only relevant for multiclass models. When it is set to a 0-based class index,
only SHAP contributions for that specific class are used.
If it is not set, SHAP importances are averaged over all classes.}
\item{approxcontrib}{passed to \code{\link{predict.xgb.Booster}} when \code{shap_contrib = NULL}.}
\item{subsample}{a random fraction of data points to use for plotting. When it is NULL,
it is set so that up to 100K data points are used.}
}
\value{
A list containing: 'data', a matrix containing sample observations
and their feature values; 'shap_contrib', a matrix containing the SHAP contribution
values for these observations.
}
\description{
Prepare data for SHAP plots. To be used in xgb.plot.shap, xgb.plot.shap.summary, etc.
Internal utility function.
}
\keyword{internal}

283
R-package/man/xgb.train.Rd
View File

@ -43,111 +43,114 @@ xgboost(
}
\arguments{
\item{params}{the list of parameters. The complete list of parameters is
available in the \href{http://xgboost.readthedocs.io/en/latest/parameter.html}{online documentation}. Below
is a shorter summary:
1. General Parameters
\itemize{
\item \code{booster} which booster to use, can be \code{gbtree} or \code{gblinear}. Default: \code{gbtree}.
available in the \href{http://xgboost.readthedocs.io/en/latest/parameter.html}{online documentation}. Below
is a shorter summary:
\enumerate{
\item General Parameters
}
2. Booster Parameters
\itemize{
\item \code{booster} which booster to use, can be \code{gbtree} or \code{gblinear}. Default: \code{gbtree}.
}
\enumerate{
\item Booster Parameters
}
2.1. Parameters for Tree Booster
\itemize{
\item{ \code{eta} control the learning rate: scale the contribution of each tree by a factor of \code{0 < eta < 1}
when it is added to the current approximation.
Used to prevent overfitting by making the boosting process more conservative.
Lower value for \code{eta} implies larger value for \code{nrounds}: low \code{eta} value means model
more robust to overfitting but slower to compute. Default: 0.3}
\item{ \code{gamma} minimum loss reduction required to make a further partition on a leaf node of the tree.
the larger, the more conservative the algorithm will be.}
\item \code{max_depth} maximum depth of a tree. Default: 6
\item{\code{min_child_weight} minimum sum of instance weight (hessian) needed in a child.
If the tree partition step results in a leaf node with the sum of instance weight less than min_child_weight,
then the building process will give up further partitioning.
In linear regression mode, this simply corresponds to minimum number of instances needed to be in each node.
The larger, the more conservative the algorithm will be. Default: 1}
\item{ \code{subsample} subsample ratio of the training instance.
Setting it to 0.5 means that xgboost randomly collected half of the data instances to grow trees
and this will prevent overfitting. It makes computation shorter (because less data to analyse).
It is advised to use this parameter with \code{eta} and increase \code{nrounds}. Default: 1}
\item \code{colsample_bytree} subsample ratio of columns when constructing each tree. Default: 1
\item \code{lambda} L2 regularization term on weights. Default: 1
\item \code{alpha} L1 regularization term on weights. (there is no L1 reg on bias because it is not important). Default: 0
\item{ \code{num_parallel_tree} Experimental parameter. number of trees to grow per round.
Useful to test Random Forest through XGBoost
(set \code{colsample_bytree < 1}, \code{subsample < 1} and \code{round = 1}) accordingly.
Default: 1}
\item{ \code{monotone_constraints} A numerical vector consists of \code{1}, \code{0} and \code{-1} with its length
equals to the number of features in the training data.
\code{1} is increasing, \code{-1} is decreasing and \code{0} is no constraint.}
\item{ \code{interaction_constraints} A list of vectors specifying feature indices of permitted interactions.
Each item of the list represents one permitted interaction where specified features are allowed to interact with each other.
Feature index values should start from \code{0} (\code{0} references the first column).
Leave argument unspecified for no interaction constraints.}
\item{ \code{eta} control the learning rate: scale the contribution of each tree by a factor of \code{0 < eta < 1}
when it is added to the current approximation.
Used to prevent overfitting by making the boosting process more conservative.
Lower value for \code{eta} implies larger value for \code{nrounds}: low \code{eta} value means model
more robust to overfitting but slower to compute. Default: 0.3}
\item{ \code{gamma} minimum loss reduction required to make a further partition on a leaf node of the tree.
the larger, the more conservative the algorithm will be.}
\item \code{max_depth} maximum depth of a tree. Default: 6
\item{\code{min_child_weight} minimum sum of instance weight (hessian) needed in a child.
If the tree partition step results in a leaf node with the sum of instance weight less than min_child_weight,
then the building process will give up further partitioning.
In linear regression mode, this simply corresponds to minimum number of instances needed to be in each node.
The larger, the more conservative the algorithm will be. Default: 1}
\item{ \code{subsample} subsample ratio of the training instance.
Setting it to 0.5 means that xgboost randomly collected half of the data instances to grow trees
and this will prevent overfitting. It makes computation shorter (because less data to analyse).
It is advised to use this parameter with \code{eta} and increase \code{nrounds}. Default: 1}
\item \code{colsample_bytree} subsample ratio of columns when constructing each tree. Default: 1
\item \code{lambda} L2 regularization term on weights. Default: 1
\item \code{alpha} L1 regularization term on weights. (there is no L1 reg on bias because it is not important). Default: 0
\item{ \code{num_parallel_tree} Experimental parameter. number of trees to grow per round.
Useful to test Random Forest through XGBoost
(set \code{colsample_bytree < 1}, \code{subsample < 1} and \code{round = 1}) accordingly.
Default: 1}
\item{ \code{monotone_constraints} A numerical vector consists of \code{1}, \code{0} and \code{-1} with its length
equals to the number of features in the training data.
\code{1} is increasing, \code{-1} is decreasing and \code{0} is no constraint.}
\item{ \code{interaction_constraints} A list of vectors specifying feature indices of permitted interactions.
Each item of the list represents one permitted interaction where specified features are allowed to interact with each other.
Feature index values should start from \code{0} (\code{0} references the first column).
Leave argument unspecified for no interaction constraints.}
}
2.2. Parameters for Linear Booster
\itemize{
\item \code{lambda} L2 regularization term on weights. Default: 0
\item \code{lambda_bias} L2 regularization term on bias. Default: 0
\item \code{alpha} L1 regularization term on weights. (there is no L1 reg on bias because it is not important). Default: 0
\item \code{lambda} L2 regularization term on weights. Default: 0
\item \code{lambda_bias} L2 regularization term on bias. Default: 0
\item \code{alpha} L1 regularization term on weights. (there is no L1 reg on bias because it is not important). Default: 0
}
\enumerate{
\item Task Parameters
}
3. Task Parameters
\itemize{
\item{ \code{objective} specify the learning task and the corresponding learning objective, users can pass a self-defined function to it.
The default objective options are below:
\itemize{
\item \code{reg:squarederror} Regression with squared loss (Default).
\item{ \code{reg:squaredlogerror}: regression with squared log loss \eqn{1/2 * (log(pred + 1) - log(label + 1))^2}.
All inputs are required to be greater than -1.
Also, see metric rmsle for possible issue with this objective.}
\item \code{reg:logistic} logistic regression.
\item \code{reg:pseudohubererror}: regression with Pseudo Huber loss, a twice differentiable alternative to absolute loss.
\item \code{binary:logistic} logistic regression for binary classification. Output probability.
\item \code{binary:logitraw} logistic regression for binary classification, output score before logistic transformation.
\item \code{binary:hinge}: hinge loss for binary classification. This makes predictions of 0 or 1, rather than producing probabilities.
\item{ \code{count:poisson}: Poisson regression for count data, output mean of Poisson distribution.
\code{max_delta_step} is set to 0.7 by default in poisson regression (used to safeguard optimization).}
\item{ \code{survival:cox}: Cox regression for right censored survival time data (negative values are considered right censored).
Note that predictions are returned on the hazard ratio scale (i.e., as HR = exp(marginal_prediction) in the proportional
hazard function \code{h(t) = h0(t) * HR)}.}
\item{ \code{survival:aft}: Accelerated failure time model for censored survival time data. See
\href{https://xgboost.readthedocs.io/en/latest/tutorials/aft_survival_analysis.html}{Survival Analysis with Accelerated Failure Time}
for details.}
\item \code{aft_loss_distribution}: Probability Density Function used by \code{survival:aft} and \code{aft-nloglik} metric.
\item{ \code{multi:softmax} set xgboost to do multiclass classification using the softmax objective.
Class is represented by a number and should be from 0 to \code{num_class - 1}.}
\item{ \code{multi:softprob} same as softmax, but prediction outputs a vector of ndata * nclass elements, which can be
further reshaped to ndata, nclass matrix. The result contains predicted probabilities of each data point belonging
to each class.}
\item \code{rank:pairwise} set xgboost to do ranking task by minimizing the pairwise loss.
\item{ \code{rank:ndcg}: Use LambdaMART to perform list-wise ranking where
\href{https://en.wikipedia.org/wiki/Discounted_cumulative_gain}{Normalized Discounted Cumulative Gain (NDCG)} is maximized.}
\item{ \code{rank:map}: Use LambdaMART to perform list-wise ranking where
\href{https://en.wikipedia.org/wiki/Evaluation_measures_(information_retrieval)#Mean_average_precision}{Mean Average Precision (MAP)}
is maximized.}
\item{ \code{reg:gamma}: gamma regression with log-link.
Output is a mean of gamma distribution.
It might be useful, e.g., for modeling insurance claims severity, or for any outcome that might be
\href{https://en.wikipedia.org/wiki/Gamma_distribution#Applications}{gamma-distributed}.}
\item{ \code{reg:tweedie}: Tweedie regression with log-link.
It might be useful, e.g., for modeling total loss in insurance, or for any outcome that might be
\href{https://en.wikipedia.org/wiki/Tweedie_distribution#Applications}{Tweedie-distributed}.}
}
}
\item \code{base_score} the initial prediction score of all instances, global bias. Default: 0.5
\item{ \code{eval_metric} evaluation metrics for validation data.
Users can pass a self-defined function to it.
Default: metric will be assigned according to objective
(rmse for regression, and error for classification, mean average precision for ranking).
List is provided in detail section.}
The default objective options are below:
\itemize{
\item \code{reg:squarederror} Regression with squared loss (Default).
\item{ \code{reg:squaredlogerror}: regression with squared log loss \eqn{1/2 * (log(pred + 1) - log(label + 1))^2}.
All inputs are required to be greater than -1.
Also, see metric rmsle for possible issue with this objective.}
\item \code{reg:logistic} logistic regression.
\item \code{reg:pseudohubererror}: regression with Pseudo Huber loss, a twice differentiable alternative to absolute loss.
\item \code{binary:logistic} logistic regression for binary classification. Output probability.
\item \code{binary:logitraw} logistic regression for binary classification, output score before logistic transformation.
\item \code{binary:hinge}: hinge loss for binary classification. This makes predictions of 0 or 1, rather than producing probabilities.
\item{ \code{count:poisson}: Poisson regression for count data, output mean of Poisson distribution.
\code{max_delta_step} is set to 0.7 by default in poisson regression (used to safeguard optimization).}
\item{ \code{survival:cox}: Cox regression for right censored survival time data (negative values are considered right censored).
Note that predictions are returned on the hazard ratio scale (i.e., as HR = exp(marginal_prediction) in the proportional
hazard function \code{h(t) = h0(t) * HR)}.}
\item{ \code{survival:aft}: Accelerated failure time model for censored survival time data. See
\href{https://xgboost.readthedocs.io/en/latest/tutorials/aft_survival_analysis.html}{Survival Analysis with Accelerated Failure Time}
for details.}
\item \code{aft_loss_distribution}: Probability Density Function used by \code{survival:aft} and \code{aft-nloglik} metric.
\item{ \code{multi:softmax} set xgboost to do multiclass classification using the softmax objective.
Class is represented by a number and should be from 0 to \code{num_class - 1}.}
\item{ \code{multi:softprob} same as softmax, but prediction outputs a vector of ndata * nclass elements, which can be
further reshaped to ndata, nclass matrix. The result contains predicted probabilities of each data point belonging
to each class.}
\item \code{rank:pairwise} set xgboost to do ranking task by minimizing the pairwise loss.
\item{ \code{rank:ndcg}: Use LambdaMART to perform list-wise ranking where
\href{https://en.wikipedia.org/wiki/Discounted_cumulative_gain}{Normalized Discounted Cumulative Gain (NDCG)} is maximized.}
\item{ \code{rank:map}: Use LambdaMART to perform list-wise ranking where
\href{https://en.wikipedia.org/wiki/Evaluation_measures_(information_retrieval)#Mean_average_precision}{Mean Average Precision (MAP)}
is maximized.}
\item{ \code{reg:gamma}: gamma regression with log-link.
Output is a mean of gamma distribution.
It might be useful, e.g., for modeling insurance claims severity, or for any outcome that might be
\href{https://en.wikipedia.org/wiki/Gamma_distribution#Applications}{gamma-distributed}.}
\item{ \code{reg:tweedie}: Tweedie regression with log-link.
It might be useful, e.g., for modeling total loss in insurance, or for any outcome that might be
\href{https://en.wikipedia.org/wiki/Tweedie_distribution#Applications}{Tweedie-distributed}.}
}
}
\item \code{base_score} the initial prediction score of all instances, global bias. Default: 0.5
\item{ \code{eval_metric} evaluation metrics for validation data.
Users can pass a self-defined function to it.
Default: metric will be assigned according to objective
(rmse for regression, and error for classification, mean average precision for ranking).
List is provided in detail section.}
}}
\item{data}{training dataset. \code{xgb.train} accepts only an \code{xgb.DMatrix} as the input.
@ -202,7 +205,12 @@ file with a previously saved model.}
\item{callbacks}{a list of callback functions to perform various task during boosting.
See \code{\link{callbacks}}. Some of the callbacks are automatically created depending on the
parameters' values. User can provide either existing or their own callback methods in order
to customize the training process.}
to customize the training process.
\if{html}{\out{<div class="sourceCode">}}\preformatted{ Note that some callbacks might try to set an evaluation log - be aware that these evaluation logs
are kept as R attributes, and thus do not get saved when using non-R serializaters like
\link{xgb.save} (but are kept when using R serializers like \link{saveRDS}).
}\if{html}{\out{</div>}}}
\item{...}{other parameters to pass to \code{params}.}
@ -216,27 +224,7 @@ This parameter is only used when input is a dense matrix.}
\item{weight}{a vector indicating the weight for each row of the input.}
}
\value{
An object of class \code{xgb.Booster} with the following elements:
\itemize{
\item \code{handle} a handle (pointer) to the xgboost model in memory.
\item \code{raw} a cached memory dump of the xgboost model saved as R's \code{raw} type.
\item \code{niter} number of boosting iterations.
\item \code{evaluation_log} evaluation history stored as a \code{data.table} with the
first column corresponding to iteration number and the rest corresponding to evaluation
metrics' values. It is created by the \code{\link{cb.evaluation.log}} callback.
\item \code{call} a function call.
\item \code{params} parameters that were passed to the xgboost library. Note that it does not
capture parameters changed by the \code{\link{cb.reset.parameters}} callback.
\item \code{callbacks} callback functions that were either automatically assigned or
explicitly passed.
\item \code{best_iteration} iteration number with the best evaluation metric value
(only available with early stopping).
\item \code{best_score} the best evaluation metric value during early stopping.
(only available with early stopping).
\item \code{feature_names} names of the training dataset features
(only when column names were defined in training data).
\item \code{nfeatures} number of features in training data.
}
An object of class \code{xgb.Booster}.
}
\description{
\code{xgb.train} is an advanced interface for training an xgboost model.
@ -258,30 +246,45 @@ when the \code{eval_metric} parameter is not provided.
User may set one or several \code{eval_metric} parameters.
Note that when using a customized metric, only this single metric can be used.
The following is the list of built-in metrics for which XGBoost provides optimized implementation:
\itemize{
\item \code{rmse} root mean square error. \url{https://en.wikipedia.org/wiki/Root_mean_square_error}
\item \code{logloss} negative log-likelihood. \url{https://en.wikipedia.org/wiki/Log-likelihood}
\item \code{mlogloss} multiclass logloss. \url{https://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html}
\item \code{error} Binary classification error rate. It is calculated as \code{(# wrong cases) / (# all cases)}.
By default, it uses the 0.5 threshold for predicted values to define negative and positive instances.
Different threshold (e.g., 0.) could be specified as "error@0."
\item \code{merror} Multiclass classification error rate. It is calculated as \code{(# wrong cases) / (# all cases)}.
\item \code{mae} Mean absolute error
\item \code{mape} Mean absolute percentage error
\item{ \code{auc} Area under the curve.
\url{https://en.wikipedia.org/wiki/Receiver_operating_characteristic#'Area_under_curve} for ranking evaluation.}
\item \code{aucpr} Area under the PR curve. \url{https://en.wikipedia.org/wiki/Precision_and_recall} for ranking evaluation.
\item \code{ndcg} Normalized Discounted Cumulative Gain (for ranking task). \url{https://en.wikipedia.org/wiki/NDCG}
}
\itemize{
\item \code{rmse} root mean square error. \url{https://en.wikipedia.org/wiki/Root_mean_square_error}
\item \code{logloss} negative log-likelihood. \url{https://en.wikipedia.org/wiki/Log-likelihood}
\item \code{mlogloss} multiclass logloss. \url{https://scikit-learn.org/stable/modules/generated/sklearn.metrics.log_loss.html}
\item \code{error} Binary classification error rate. It is calculated as \code{(# wrong cases) / (# all cases)}.
By default, it uses the 0.5 threshold for predicted values to define negative and positive instances.
Different threshold (e.g., 0.) could be specified as "error@0."
\item \code{merror} Multiclass classification error rate. It is calculated as \code{(# wrong cases) / (# all cases)}.
\item \code{mae} Mean absolute error
\item \code{mape} Mean absolute percentage error
\item{ \code{auc} Area under the curve.
\url{https://en.wikipedia.org/wiki/Receiver_operating_characteristic#'Area_under_curve} for ranking evaluation.}
\item \code{aucpr} Area under the PR curve. \url{https://en.wikipedia.org/wiki/Precision_and_recall} for ranking evaluation.
\item \code{ndcg} Normalized Discounted Cumulative Gain (for ranking task). \url{https://en.wikipedia.org/wiki/NDCG}
}
The following callbacks are automatically created when certain parameters are set:
\itemize{
\item \code{cb.print.evaluation} is turned on when \code{verbose > 0};
and the \code{print_every_n} parameter is passed to it.
\item \code{cb.evaluation.log} is on when \code{watchlist} is present.
\item \code{cb.early.stop}: when \code{early_stopping_rounds} is set.
\item \code{cb.save.model}: when \code{save_period > 0} is set.
\item \code{cb.print.evaluation} is turned on when \code{verbose > 0};
and the \code{print_every_n} parameter is passed to it.
\item \code{cb.evaluation.log} is on when \code{watchlist} is present.
\item \code{cb.early.stop}: when \code{early_stopping_rounds} is set.
\item \code{cb.save.model}: when \code{save_period > 0} is set.
}
Note that objects of type \code{xgb.Booster} as returned by this function behave a bit differently
from typical R objects (it's an 'altrep' list class), and it makes a separation between
internal booster attributes (restricted to jsonifyable data), accessed through \link{xgb.attr}
and shared between interfaces through serialization functions like \link{xgb.save}; and
R-specific attributes, accessed through \link{attributes} and \link{attr}, which are otherwise
only used in the R interface, only kept when using R's serializers like \link{saveRDS}, and
not anyhow used by functions like \link{predict.xgb.Booster}.
Be aware that one such R attribute that is automatically added is \code{params} - this attribute
is assigned from the \code{params} argument to this function, and is only meant to serve as a
reference for what went into the booster, but is not used in other methods that take a booster
object - so for example, changing the booster's configuration requires calling \verb{xgb.config<-}
or 'xgb.parameters<-', while simply modifying \verb{attributes(model)$params$<...>} will have no
effect elsewhere.
}
\examples{
data(agaricus.train, package='xgboost')
@ -300,9 +303,9 @@ dtest <- with(
watchlist <- list(train = dtrain, eval = dtest)
## A simple xgb.train example:
param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
param <- list(max_depth = 2, eta = 1, nthread = nthread,
objective = "binary:logistic", eval_metric = "auc")
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0)
## An xgb.train example where custom objective and evaluation metric are
## used:
@ -321,13 +324,13 @@ evalerror <- function(preds, dtrain) {
# These functions could be used by passing them either:
# as 'objective' and 'eval_metric' parameters in the params list:
param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
param <- list(max_depth = 2, eta = 1, nthread = nthread,
objective = logregobj, eval_metric = evalerror)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0)
# or through the ... arguments:
param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist,
param <- list(max_depth = 2, eta = 1, nthread = nthread)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
objective = logregobj, eval_metric = evalerror)
# or as dedicated 'obj' and 'feval' parameters of xgb.train:
@ -336,10 +339,10 @@ bst <- xgb.train(param, dtrain, nrounds = 2, watchlist,
## An xgb.train example of using variable learning rates at each iteration:
param <- list(max_depth = 2, eta = 1, verbose = 0, nthread = nthread,
param <- list(max_depth = 2, eta = 1, nthread = nthread,
objective = "binary:logistic", eval_metric = "auc")
my_etas <- list(eta = c(0.5, 0.1))
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist,
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
callbacks = list(cb.reset.parameters(my_etas)))
## Early stopping:

21
R-package/man/xgb.unserialize.Rd
View File

@ -1,21 +0,0 @@
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/xgb.unserialize.R
\name{xgb.unserialize}
\alias{xgb.unserialize}
\title{Load the instance back from \code{\link{xgb.serialize}}}
\usage{
xgb.unserialize(buffer, handle = NULL)
}
\arguments{
\item{buffer}{the buffer containing booster instance saved by \code{\link{xgb.serialize}}}
\item{handle}{An \code{xgb.Booster.handle} object which will be overwritten with
the new deserialized object. Must be a null handle (e.g. when loading the model through
`readRDS`). If not provided, a new handle will be created.}
}
\value{
An \code{xgb.Booster.handle} object.
}
\description{
Load the instance back from \code{\link{xgb.serialize}}
}

1
R-package/src/Makevars.in
View File

@ -63,6 +63,7 @@ OBJECTS= \
$(PKGROOT)/src/gbm/gblinear.o \
$(PKGROOT)/src/gbm/gblinear_model.o \
$(PKGROOT)/src/data/adapter.o \
$(PKGROOT)/src/data/array_interface.o \
$(PKGROOT)/src/data/simple_dmatrix.o \
$(PKGROOT)/src/data/data.o \
$(PKGROOT)/src/data/sparse_page_raw_format.o \

1
R-package/src/Makevars.win
View File

@ -63,6 +63,7 @@ OBJECTS= \
$(PKGROOT)/src/gbm/gblinear.o \
$(PKGROOT)/src/gbm/gblinear_model.o \
$(PKGROOT)/src/data/adapter.o \
$(PKGROOT)/src/data/array_interface.o \
$(PKGROOT)/src/data/simple_dmatrix.o \
$(PKGROOT)/src/data/data.o \
$(PKGROOT)/src/data/sparse_page_raw_format.o \

34
R-package/src/init.c
View File

@ -15,9 +15,16 @@ Check these declarations against the C/Fortran source code.
*/
/* .Call calls */
extern void XGBInitializeAltrepClass_R(DllInfo *info);
extern SEXP XGDuplicate_R(SEXP);
extern SEXP XGPointerEqComparison_R(SEXP, SEXP);
extern SEXP XGBoosterTrainOneIter_R(SEXP, SEXP, SEXP, SEXP, SEXP);
extern SEXP XGBoosterCreate_R(SEXP);
extern SEXP XGBoosterCreateInEmptyObj_R(SEXP, SEXP);
extern SEXP XGBoosterCopyInfoFromDMatrix_R(SEXP, SEXP);
extern SEXP XGBoosterSetStrFeatureInfo_R(SEXP, SEXP, SEXP);
extern SEXP XGBoosterGetStrFeatureInfo_R(SEXP, SEXP);
extern SEXP XGBoosterBoostedRounds_R(SEXP);
extern SEXP XGBoosterGetNumFeature_R(SEXP);
extern SEXP XGBoosterDumpModel_R(SEXP, SEXP, SEXP, SEXP);
extern SEXP XGBoosterEvalOneIter_R(SEXP, SEXP, SEXP, SEXP);
extern SEXP XGBoosterGetAttrNames_R(SEXP);
@ -39,10 +46,15 @@ extern SEXP XGDMatrixCreateFromCSC_R(SEXP, SEXP, SEXP, SEXP, SEXP, SEXP);
extern SEXP XGDMatrixCreateFromCSR_R(SEXP, SEXP, SEXP, SEXP, SEXP, SEXP);
extern SEXP XGDMatrixCreateFromFile_R(SEXP, SEXP);
extern SEXP XGDMatrixCreateFromMat_R(SEXP, SEXP, SEXP);
extern SEXP XGDMatrixGetInfo_R(SEXP, SEXP);
extern SEXP XGDMatrixGetFloatInfo_R(SEXP, SEXP);
extern SEXP XGDMatrixGetUIntInfo_R(SEXP, SEXP);
extern SEXP XGDMatrixCreateFromDF_R(SEXP, SEXP, SEXP);
extern SEXP XGDMatrixGetStrFeatureInfo_R(SEXP, SEXP);
extern SEXP XGDMatrixNumCol_R(SEXP);
extern SEXP XGDMatrixNumRow_R(SEXP);
extern SEXP XGDMatrixGetQuantileCut_R(SEXP);
extern SEXP XGDMatrixNumNonMissing_R(SEXP);
extern SEXP XGDMatrixGetDataAsCSR_R(SEXP);
extern SEXP XGDMatrixSaveBinary_R(SEXP, SEXP, SEXP);
extern SEXP XGDMatrixSetInfo_R(SEXP, SEXP, SEXP);
extern SEXP XGDMatrixSetStrFeatureInfo_R(SEXP, SEXP, SEXP);
@ -52,9 +64,15 @@ extern SEXP XGBGetGlobalConfig_R(void);
extern SEXP XGBoosterFeatureScore_R(SEXP, SEXP);
static const R_CallMethodDef CallEntries[] = {
{"XGBoosterBoostOneIter_R", (DL_FUNC) &XGBoosterTrainOneIter_R, 5},
{"XGDuplicate_R", (DL_FUNC) &XGDuplicate_R, 1},
{"XGPointerEqComparison_R", (DL_FUNC) &XGPointerEqComparison_R, 2},
{"XGBoosterTrainOneIter_R", (DL_FUNC) &XGBoosterTrainOneIter_R, 5},
{"XGBoosterCreate_R", (DL_FUNC) &XGBoosterCreate_R, 1},
{"XGBoosterCreateInEmptyObj_R", (DL_FUNC) &XGBoosterCreateInEmptyObj_R, 2},
{"XGBoosterCopyInfoFromDMatrix_R", (DL_FUNC) &XGBoosterCopyInfoFromDMatrix_R, 2},
{"XGBoosterSetStrFeatureInfo_R",(DL_FUNC) &XGBoosterSetStrFeatureInfo_R,3}, // NOLINT
{"XGBoosterGetStrFeatureInfo_R",(DL_FUNC) &XGBoosterGetStrFeatureInfo_R,2}, // NOLINT
{"XGBoosterBoostedRounds_R", (DL_FUNC) &XGBoosterBoostedRounds_R, 1},
{"XGBoosterGetNumFeature_R", (DL_FUNC) &XGBoosterGetNumFeature_R, 1},
{"XGBoosterDumpModel_R", (DL_FUNC) &XGBoosterDumpModel_R, 4},
{"XGBoosterEvalOneIter_R", (DL_FUNC) &XGBoosterEvalOneIter_R, 4},
{"XGBoosterGetAttrNames_R", (DL_FUNC) &XGBoosterGetAttrNames_R, 1},
@ -76,10 +94,15 @@ static const R_CallMethodDef CallEntries[] = {
{"XGDMatrixCreateFromCSR_R", (DL_FUNC) &XGDMatrixCreateFromCSR_R, 6},
{"XGDMatrixCreateFromFile_R", (DL_FUNC) &XGDMatrixCreateFromFile_R, 2},
{"XGDMatrixCreateFromMat_R", (DL_FUNC) &XGDMatrixCreateFromMat_R, 3},
{"XGDMatrixGetInfo_R", (DL_FUNC) &XGDMatrixGetInfo_R, 2},
{"XGDMatrixGetFloatInfo_R", (DL_FUNC) &XGDMatrixGetFloatInfo_R, 2},
{"XGDMatrixGetUIntInfo_R", (DL_FUNC) &XGDMatrixGetUIntInfo_R, 2},
{"XGDMatrixCreateFromDF_R", (DL_FUNC) &XGDMatrixCreateFromDF_R, 3},
{"XGDMatrixGetStrFeatureInfo_R", (DL_FUNC) &XGDMatrixGetStrFeatureInfo_R, 2},
{"XGDMatrixNumCol_R", (DL_FUNC) &XGDMatrixNumCol_R, 1},
{"XGDMatrixNumRow_R", (DL_FUNC) &XGDMatrixNumRow_R, 1},
{"XGDMatrixGetQuantileCut_R", (DL_FUNC) &XGDMatrixGetQuantileCut_R, 1},
{"XGDMatrixNumNonMissing_R", (DL_FUNC) &XGDMatrixNumNonMissing_R, 1},
{"XGDMatrixGetDataAsCSR_R", (DL_FUNC) &XGDMatrixGetDataAsCSR_R, 1},
{"XGDMatrixSaveBinary_R", (DL_FUNC) &XGDMatrixSaveBinary_R, 3},
{"XGDMatrixSetInfo_R", (DL_FUNC) &XGDMatrixSetInfo_R, 3},
{"XGDMatrixSetStrFeatureInfo_R", (DL_FUNC) &XGDMatrixSetStrFeatureInfo_R, 3},
@ -96,4 +119,5 @@ __declspec(dllexport)
void attribute_visible R_init_xgboost(DllInfo *dll) {
R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
R_useDynamicSymbols(dll, FALSE);
XGBInitializeAltrepClass_R(dll);
}

551
R-package/src/xgboost_R.cc
View File

@ -1,5 +1,5 @@
/**
* Copyright 2014-2023 by XGBoost Contributors
* Copyright 2014-2024, XGBoost Contributors
*/
#include <dmlc/common.h>
#include <dmlc/omp.h>
@ -8,9 +8,12 @@
#include <xgboost/data.h>
#include <xgboost/logging.h>
#include <algorithm>
#include <cmath>
#include <cstdint>
#include <cstdio>
#include <cstring>
#include <limits>
#include <sstream>
#include <string>
#include <utility>
@ -19,14 +22,14 @@
#include "../../src/c_api/c_api_error.h"
#include "../../src/c_api/c_api_utils.h" // MakeSparseFromPtr
#include "../../src/common/threading_utils.h"
#include "../../src/data/array_interface.h" // for ArrayInterface
#include "./xgboost_R.h" // Must follow other includes.
namespace {
struct ErrorWithUnwind : public std::exception {};
void ThrowExceptionFromRError(void *unused, Rboolean jump) {
void ThrowExceptionFromRError(void *, Rboolean jump) {
if (jump) {
throw ErrorWithUnwind();
}
@ -48,8 +51,35 @@ SEXP SafeMkChar(const char *c_str, SEXP continuation_token) {
continuation_token);
}
SEXP WrappedAllocReal(void *void_ptr) {
size_t *size = static_cast<size_t*>(void_ptr);
return Rf_allocVector(REALSXP, *size);
}
SEXP SafeAllocReal(size_t size, SEXP continuation_token) {
return R_UnwindProtect(
WrappedAllocReal, static_cast<void*>(&size),
ThrowExceptionFromRError, nullptr,
continuation_token);
}
SEXP WrappedAllocInteger(void *void_ptr) {
size_t *size = static_cast<size_t*>(void_ptr);
return Rf_allocVector(INTSXP, *size);
}
SEXP SafeAllocInteger(size_t size, SEXP continuation_token) {
return R_UnwindProtect(
WrappedAllocInteger, static_cast<void*>(&size),
ThrowExceptionFromRError, nullptr,
continuation_token);
}
[[nodiscard]] std::string MakeArrayInterfaceFromRMat(SEXP R_mat) {
SEXP mat_dims = Rf_getAttrib(R_mat, R_DimSymbol);
if (Rf_xlength(mat_dims) > 2) {
LOG(FATAL) << "Passed input array with more than two dimensions, which is not supported.";
}
const int *ptr_mat_dims = INTEGER(mat_dims);
// Lambda for type dispatch.
@ -132,45 +162,116 @@ SEXP SafeMkChar(const char *c_str, SEXP continuation_token) {
jconfig["nthread"] = Rf_asInteger(n_threads);
return Json::Dump(jconfig);
}
// Allocate a R vector and copy an array interface encoded object to it.
[[nodiscard]] SEXP CopyArrayToR(const char *array_str, SEXP ctoken) {
xgboost::ArrayInterface<1> array{xgboost::StringView{array_str}};
// R supports only int and double.
bool is_int_type =
xgboost::DispatchDType(array.type, [](auto t) { return std::is_integral_v<decltype(t)>; });
bool is_float = xgboost::DispatchDType(
array.type, [](auto v) { return std::is_floating_point_v<decltype(v)>; });
CHECK(is_int_type || is_float) << "Internal error: Invalid DType.";
CHECK(array.is_contiguous) << "Internal error: Return by XGBoost should be contiguous";
// Note: the only case in which this will receive an integer type is
// for the 'indptr' part of the quantile cut outputs, which comes
// in sorted order, so the last element contains the maximum value.
bool fits_into_C_int = xgboost::DispatchDType(array.type, [&](auto t) {
using T = decltype(t);
if (!std::is_integral_v<decltype(t)>) {
return false;
}
auto ptr = static_cast<T const *>(array.data);
T last_elt = ptr[array.n - 1];
if (last_elt < 0) {
last_elt = -last_elt; // no std::abs overload for all possible types
}
return last_elt <= std::numeric_limits<int>::max();
});
bool use_int = is_int_type && fits_into_C_int;
// Allocate memory in R
SEXP out =
Rf_protect(use_int ? SafeAllocInteger(array.n, ctoken) : SafeAllocReal(array.n, ctoken));
xgboost::DispatchDType(array.type, [&](auto t) {
using T = decltype(t);
auto in_ptr = static_cast<T const *>(array.data);
if (use_int) {
auto out_ptr = INTEGER(out);
std::copy_n(in_ptr, array.n, out_ptr);
} else {
auto out_ptr = REAL(out);
std::copy_n(in_ptr, array.n, out_ptr);
}
});
Rf_unprotect(1);
return out;
}
} // namespace
struct RRNGStateController {
RRNGStateController() {
GetRNGstate();
}
~RRNGStateController() {
PutRNGstate();
}
};
/*!
* \brief macro to annotate begin of api
*/
#define R_API_BEGIN() \
GetRNGstate(); \
try {
try { \
RRNGStateController rng_controller{};
/* Note: an R error triggers a long jump, hence all C++ objects that
allocated memory through non-R allocators, including the exception
object, need to be destructed before triggering the R error.
In order to preserve the error message, it gets copied to a temporary
buffer, and the R error section is reached through a 'goto' statement
that bypasses usual function control flow. */
char cpp_ex_msg[512];
/*!
* \brief macro to annotate end of api
*/
#define R_API_END() \
} catch(dmlc::Error& e) { \
PutRNGstate(); \
error(e.what()); \
} catch(std::exception &e) { \
std::strncpy(cpp_ex_msg, e.what(), 512); \
goto throw_cpp_ex_as_R_err; \
} \
PutRNGstate();
if (false) { \
throw_cpp_ex_as_R_err: \
Rf_error("%s", cpp_ex_msg); \
}
/*!
* \brief macro to check the call.
/**
* @brief Macro for checking XGBoost return code.
*/
#define CHECK_CALL(x) \
if ((x) != 0) { \
error(XGBGetLastError()); \
#define CHECK_CALL(__rc) \
if ((__rc) != 0) { \
Rf_error("%s", XGBGetLastError()); \
}
using dmlc::BeginPtr;
XGB_DLL SEXP XGCheckNullPtr_R(SEXP handle) {
return ScalarLogical(R_ExternalPtrAddr(handle) == NULL);
return Rf_ScalarLogical(R_ExternalPtrAddr(handle) == nullptr);
}
XGB_DLL void _DMatrixFinalizer(SEXP ext) {
namespace {
void _DMatrixFinalizer(SEXP ext) {
R_API_BEGIN();
if (R_ExternalPtrAddr(ext) == NULL) return;
CHECK_CALL(XGDMatrixFree(R_ExternalPtrAddr(ext)));
R_ClearExternalPtr(ext);
R_API_END();
}
} /* namespace */
XGB_DLL SEXP XGBSetGlobalConfig_R(SEXP json_str) {
R_API_BEGIN();
@ -219,6 +320,69 @@ XGB_DLL SEXP XGDMatrixCreateFromMat_R(SEXP mat, SEXP missing, SEXP n_threads) {
return ret;
}
XGB_DLL SEXP XGDMatrixCreateFromDF_R(SEXP df, SEXP missing, SEXP n_threads) {
SEXP ret = Rf_protect(R_MakeExternalPtr(nullptr, R_NilValue, R_NilValue));
R_API_BEGIN();
DMatrixHandle handle;
auto make_vec = [&](auto const *ptr, std::int32_t len) {
auto v = xgboost::linalg::MakeVec(ptr, len);
return xgboost::linalg::ArrayInterface(v);
};
std::int32_t rc{0};
{
using xgboost::Json;
auto n_features = Rf_xlength(df);
std::vector<Json> array(n_features);
CHECK_GT(n_features, 0);
auto len = Rf_xlength(VECTOR_ELT(df, 0));
// The `data.frame` in R actually converts all data into numeric. The other type
// handlers here are not used. At the moment they are kept as a reference for when we
// can avoid making data copies during transformation.
for (decltype(n_features) i = 0; i < n_features; ++i) {
switch (TYPEOF(VECTOR_ELT(df, i))) {
case INTSXP: {
auto const *ptr = INTEGER(VECTOR_ELT(df, i));
array[i] = make_vec(ptr, len);
break;
}
case REALSXP: {
auto const *ptr = REAL(VECTOR_ELT(df, i));
array[i] = make_vec(ptr, len);
break;
}
case LGLSXP: {
auto const *ptr = LOGICAL(VECTOR_ELT(df, i));
array[i] = make_vec(ptr, len);
break;
}
default: {
LOG(FATAL) << "data.frame has unsupported type.";
}
}
}
Json jinterface{std::move(array)};
auto sinterface = Json::Dump(jinterface);
Json jconfig{xgboost::Object{}};
jconfig["missing"] = asReal(missing);
jconfig["nthread"] = asInteger(n_threads);
auto sconfig = Json::Dump(jconfig);
rc = XGDMatrixCreateFromColumnar(sinterface.c_str(), sconfig.c_str(), &handle);
}
CHECK_CALL(rc);
R_SetExternalPtrAddr(ret, handle);
R_RegisterCFinalizerEx(ret, _DMatrixFinalizer, TRUE);
R_API_END();
Rf_unprotect(1);
return ret;
}
namespace {
void CreateFromSparse(SEXP indptr, SEXP indices, SEXP data, std::string *indptr_str,
std::string *indices_str, std::string *data_str) {
@ -294,6 +458,7 @@ XGB_DLL SEXP XGDMatrixCreateFromCSR_R(SEXP indptr, SEXP indices, SEXP data, SEXP
res_code = XGDMatrixCreateFromCSR(sindptr.c_str(), sindices.c_str(), sdata.c_str(), ncol,
config.c_str(), &handle);
}
CHECK_CALL(res_code);
R_SetExternalPtrAddr(ret, handle);
R_RegisterCFinalizerEx(ret, _DMatrixFinalizer, TRUE);
R_API_END();
@ -342,9 +507,11 @@ XGB_DLL SEXP XGDMatrixSaveBinary_R(SEXP handle, SEXP fname, SEXP silent) {
XGB_DLL SEXP XGDMatrixSetInfo_R(SEXP handle, SEXP field, SEXP array) {
R_API_BEGIN();
SEXP field_ = PROTECT(Rf_asChar(field));
SEXP arr_dim = Rf_getAttrib(array, R_DimSymbol);
int res_code;
{
const std::string array_str = MakeArrayInterfaceFromRVector(array);
const std::string array_str = Rf_isNull(arr_dim)?
MakeArrayInterfaceFromRVector(array) : MakeArrayInterfaceFromRMat(array);
res_code = XGDMatrixSetInfoFromInterface(
R_ExternalPtrAddr(handle), CHAR(field_), array_str.c_str());
}
@ -362,8 +529,14 @@ XGB_DLL SEXP XGDMatrixSetStrFeatureInfo_R(SEXP handle, SEXP field, SEXP array) {
}
SEXP str_info_holder = PROTECT(Rf_allocVector(VECSXP, len));
for (size_t i = 0; i < len; ++i) {
SET_VECTOR_ELT(str_info_holder, i, Rf_asChar(VECTOR_ELT(array, i)));
if (TYPEOF(array) == STRSXP) {
for (size_t i = 0; i < len; ++i) {
SET_VECTOR_ELT(str_info_holder, i, STRING_ELT(array, i));
}
} else {
for (size_t i = 0; i < len; ++i) {
SET_VECTOR_ELT(str_info_holder, i, Rf_asChar(VECTOR_ELT(array, i)));
}
}
SEXP field_ = PROTECT(Rf_asChar(field));
@ -407,17 +580,27 @@ XGB_DLL SEXP XGDMatrixGetStrFeatureInfo_R(SEXP handle, SEXP field) {
return ret;
}
XGB_DLL SEXP XGDMatrixGetInfo_R(SEXP handle, SEXP field) {
XGB_DLL SEXP XGDMatrixGetFloatInfo_R(SEXP handle, SEXP field) {
SEXP ret;
R_API_BEGIN();
bst_ulong olen;
const float *res;
CHECK_CALL(XGDMatrixGetFloatInfo(R_ExternalPtrAddr(handle), CHAR(asChar(field)), &olen, &res));
ret = PROTECT(allocVector(REALSXP, olen));
double *ret_ = REAL(ret);
for (size_t i = 0; i < olen; ++i) {
ret_[i] = res[i];
}
std::copy(res, res + olen, REAL(ret));
R_API_END();
UNPROTECT(1);
return ret;
}
XGB_DLL SEXP XGDMatrixGetUIntInfo_R(SEXP handle, SEXP field) {
SEXP ret;
R_API_BEGIN();
bst_ulong olen;
const unsigned *res;
CHECK_CALL(XGDMatrixGetUIntInfo(R_ExternalPtrAddr(handle), CHAR(asChar(field)), &olen, &res));
ret = PROTECT(allocVector(INTSXP, olen));
std::copy(res, res + olen, INTEGER(ret));
R_API_END();
UNPROTECT(1);
return ret;
@ -439,15 +622,210 @@ XGB_DLL SEXP XGDMatrixNumCol_R(SEXP handle) {
return ScalarInteger(static_cast<int>(ncol));
}
XGB_DLL SEXP XGDuplicate_R(SEXP obj) {
return Rf_duplicate(obj);
}
XGB_DLL SEXP XGPointerEqComparison_R(SEXP obj1, SEXP obj2) {
return Rf_ScalarLogical(R_ExternalPtrAddr(obj1) == R_ExternalPtrAddr(obj2));
}
XGB_DLL SEXP XGDMatrixGetQuantileCut_R(SEXP handle) {
const char *out_names[] = {"indptr", "data", ""};
SEXP continuation_token = Rf_protect(R_MakeUnwindCont());
SEXP out = Rf_protect(Rf_mkNamed(VECSXP, out_names));
R_API_BEGIN();
const char *out_indptr;
const char *out_data;
CHECK_CALL(XGDMatrixGetQuantileCut(R_ExternalPtrAddr(handle), "{}", &out_indptr, &out_data));
try {
SET_VECTOR_ELT(out, 0, CopyArrayToR(out_indptr, continuation_token));
SET_VECTOR_ELT(out, 1, CopyArrayToR(out_data, continuation_token));
} catch (ErrorWithUnwind &e) {
R_ContinueUnwind(continuation_token);
}
R_API_END();
Rf_unprotect(2);
return out;
}
XGB_DLL SEXP XGDMatrixNumNonMissing_R(SEXP handle) {
SEXP out = Rf_protect(Rf_allocVector(REALSXP, 1));
R_API_BEGIN();
bst_ulong out_;
CHECK_CALL(XGDMatrixNumNonMissing(R_ExternalPtrAddr(handle), &out_));
REAL(out)[0] = static_cast<double>(out_);
R_API_END();
Rf_unprotect(1);
return out;
}
XGB_DLL SEXP XGDMatrixGetDataAsCSR_R(SEXP handle) {
const char *out_names[] = {"indptr", "indices", "data", "ncols", ""};
SEXP out = Rf_protect(Rf_mkNamed(VECSXP, out_names));
R_API_BEGIN();
bst_ulong nrows, ncols, nnz;
CHECK_CALL(XGDMatrixNumRow(R_ExternalPtrAddr(handle), &nrows));
CHECK_CALL(XGDMatrixNumCol(R_ExternalPtrAddr(handle), &ncols));
CHECK_CALL(XGDMatrixNumNonMissing(R_ExternalPtrAddr(handle), &nnz));
if (std::max(nrows, ncols) > std::numeric_limits<int>::max()) {
Rf_error("%s", "Error: resulting DMatrix data does not fit into R 'dgRMatrix'.");
}
SET_VECTOR_ELT(out, 0, Rf_allocVector(INTSXP, nrows + 1));
SET_VECTOR_ELT(out, 1, Rf_allocVector(INTSXP, nnz));
SET_VECTOR_ELT(out, 2, Rf_allocVector(REALSXP, nnz));
SET_VECTOR_ELT(out, 3, Rf_ScalarInteger(ncols));
std::unique_ptr<bst_ulong[]> indptr(new bst_ulong[nrows + 1]);
std::unique_ptr<unsigned[]> indices(new unsigned[nnz]);
std::unique_ptr<float[]> data(new float[nnz]);
CHECK_CALL(XGDMatrixGetDataAsCSR(R_ExternalPtrAddr(handle),
"{}",
indptr.get(),
indices.get(),
data.get()));
std::copy(indptr.get(), indptr.get() + nrows + 1, INTEGER(VECTOR_ELT(out, 0)));
std::copy(indices.get(), indices.get() + nnz, INTEGER(VECTOR_ELT(out, 1)));
std::copy(data.get(), data.get() + nnz, REAL(VECTOR_ELT(out, 2)));
R_API_END();
Rf_unprotect(1);
return out;
}
// functions related to booster
void _BoosterFinalizer(SEXP ext) {
if (R_ExternalPtrAddr(ext) == NULL) return;
CHECK_CALL(XGBoosterFree(R_ExternalPtrAddr(ext)));
R_ClearExternalPtr(ext);
namespace {
void _BoosterFinalizer(SEXP R_ptr) {
if (R_ExternalPtrAddr(R_ptr) == NULL) return;
CHECK_CALL(XGBoosterFree(R_ExternalPtrAddr(R_ptr)));
R_ClearExternalPtr(R_ptr);
}
/* Booster is represented as an altrep list with one element which
corresponds to an 'externalptr' holding the C object, forbidding
modification by not implementing setters, and adding custom serialization. */
R_altrep_class_t XGBAltrepPointerClass;
R_xlen_t XGBAltrepPointerLength_R(SEXP R_altrepped_obj) {
return 1;
}
SEXP XGBAltrepPointerGetElt_R(SEXP R_altrepped_obj, R_xlen_t idx) {
return R_altrep_data1(R_altrepped_obj);
}
SEXP XGBMakeEmptyAltrep() {
SEXP class_name = Rf_protect(Rf_mkString("xgb.Booster"));
SEXP elt_names = Rf_protect(Rf_mkString("ptr"));
SEXP R_ptr = Rf_protect(R_MakeExternalPtr(nullptr, R_NilValue, R_NilValue));
SEXP R_altrepped_obj = Rf_protect(R_new_altrep(XGBAltrepPointerClass, R_ptr, R_NilValue));
Rf_setAttrib(R_altrepped_obj, R_NamesSymbol, elt_names);
Rf_setAttrib(R_altrepped_obj, R_ClassSymbol, class_name);
Rf_unprotect(4);
return R_altrepped_obj;
}
/* Note: the idea for separating this function from the one above is to be
able to trigger all R allocations first before doing non-R allocations. */
void XGBAltrepSetPointer(SEXP R_altrepped_obj, BoosterHandle handle) {
SEXP R_ptr = R_altrep_data1(R_altrepped_obj);
R_SetExternalPtrAddr(R_ptr, handle);
R_RegisterCFinalizerEx(R_ptr, _BoosterFinalizer, TRUE);
}
SEXP XGBAltrepSerializer_R(SEXP R_altrepped_obj) {
R_API_BEGIN();
BoosterHandle handle = R_ExternalPtrAddr(R_altrep_data1(R_altrepped_obj));
char const *serialized_bytes;
bst_ulong serialized_length;
CHECK_CALL(XGBoosterSerializeToBuffer(
handle, &serialized_length, &serialized_bytes));
SEXP R_state = Rf_protect(Rf_allocVector(RAWSXP, serialized_length));
if (serialized_length != 0) {
std::memcpy(RAW(R_state), serialized_bytes, serialized_length);
}
Rf_unprotect(1);
return R_state;
R_API_END();
return R_NilValue; /* <- should not be reached */
}
SEXP XGBAltrepDeserializer_R(SEXP unused, SEXP R_state) {
SEXP R_altrepped_obj = Rf_protect(XGBMakeEmptyAltrep());
R_API_BEGIN();
BoosterHandle handle = nullptr;
CHECK_CALL(XGBoosterCreate(nullptr, 0, &handle));
int res_code = XGBoosterUnserializeFromBuffer(handle,
RAW(R_state),
Rf_xlength(R_state));
if (res_code != 0) {
XGBoosterFree(handle);
}
CHECK_CALL(res_code);
XGBAltrepSetPointer(R_altrepped_obj, handle);
R_API_END();
Rf_unprotect(1);
return R_altrepped_obj;
}
// https://purrple.cat/blog/2018/10/14/altrep-and-cpp/
Rboolean XGBAltrepInspector_R(
SEXP x, int pre, int deep, int pvec,
void (*inspect_subtree)(SEXP, int, int, int)) {
Rprintf("Altrepped external pointer [address:%p]\n",
R_ExternalPtrAddr(R_altrep_data1(x)));
return TRUE;
}
SEXP XGBAltrepDuplicate_R(SEXP R_altrepped_obj, Rboolean deep) {
R_API_BEGIN();
if (!deep) {
SEXP out = Rf_protect(XGBMakeEmptyAltrep());
R_set_altrep_data1(out, R_altrep_data1(R_altrepped_obj));
Rf_unprotect(1);
return out;
} else {
SEXP out = Rf_protect(XGBMakeEmptyAltrep());
char const *serialized_bytes;
bst_ulong serialized_length;
CHECK_CALL(XGBoosterSerializeToBuffer(
R_ExternalPtrAddr(R_altrep_data1(R_altrepped_obj)),
&serialized_length, &serialized_bytes));
BoosterHandle new_handle = nullptr;
CHECK_CALL(XGBoosterCreate(nullptr, 0, &new_handle));
int res_code = XGBoosterUnserializeFromBuffer(new_handle,
serialized_bytes,
serialized_length);
if (res_code != 0) {
XGBoosterFree(new_handle);
}
CHECK_CALL(res_code);
XGBAltrepSetPointer(out, new_handle);
Rf_unprotect(1);
return out;
}
R_API_END();
return R_NilValue; /* <- should not be reached */
}
} /* namespace */
XGB_DLL void XGBInitializeAltrepClass_R(DllInfo *dll) {
XGBAltrepPointerClass = R_make_altlist_class("XGBAltrepPointerClass", "xgboost", dll);
R_set_altrep_Length_method(XGBAltrepPointerClass, XGBAltrepPointerLength_R);
R_set_altlist_Elt_method(XGBAltrepPointerClass, XGBAltrepPointerGetElt_R);
R_set_altrep_Inspect_method(XGBAltrepPointerClass, XGBAltrepInspector_R);
R_set_altrep_Serialized_state_method(XGBAltrepPointerClass, XGBAltrepSerializer_R);
R_set_altrep_Unserialize_method(XGBAltrepPointerClass, XGBAltrepDeserializer_R);
R_set_altrep_Duplicate_method(XGBAltrepPointerClass, XGBAltrepDuplicate_R);
}
XGB_DLL SEXP XGBoosterCreate_R(SEXP dmats) {
SEXP ret = PROTECT(R_MakeExternalPtr(nullptr, R_NilValue, R_NilValue));
SEXP out = Rf_protect(XGBMakeEmptyAltrep());
R_API_BEGIN();
R_xlen_t len = Rf_xlength(dmats);
BoosterHandle handle;
@ -461,33 +839,104 @@ XGB_DLL SEXP XGBoosterCreate_R(SEXP dmats) {
res_code = XGBoosterCreate(BeginPtr(dvec), dvec.size(), &handle);
}
CHECK_CALL(res_code);
R_SetExternalPtrAddr(ret, handle);
R_RegisterCFinalizerEx(ret, _BoosterFinalizer, TRUE);
XGBAltrepSetPointer(out, handle);
R_API_END();
UNPROTECT(1);
return ret;
Rf_unprotect(1);
return out;
}
XGB_DLL SEXP XGBoosterCreateInEmptyObj_R(SEXP dmats, SEXP R_handle) {
XGB_DLL SEXP XGBoosterCopyInfoFromDMatrix_R(SEXP booster, SEXP dmat) {
R_API_BEGIN();
R_xlen_t len = Rf_xlength(dmats);
BoosterHandle handle;
char const **feature_names;
bst_ulong len_feature_names = 0;
CHECK_CALL(XGDMatrixGetStrFeatureInfo(R_ExternalPtrAddr(dmat),
"feature_name",
&len_feature_names,
&feature_names));
if (len_feature_names) {
CHECK_CALL(XGBoosterSetStrFeatureInfo(R_ExternalPtrAddr(booster),
"feature_name",
feature_names,
len_feature_names));
}
char const **feature_types;
bst_ulong len_feature_types = 0;
CHECK_CALL(XGDMatrixGetStrFeatureInfo(R_ExternalPtrAddr(dmat),
"feature_type",
&len_feature_types,
&feature_types));
if (len_feature_types) {
CHECK_CALL(XGBoosterSetStrFeatureInfo(R_ExternalPtrAddr(booster),
"feature_type",
feature_types,
len_feature_types));
}
R_API_END();
return R_NilValue;
}
XGB_DLL SEXP XGBoosterSetStrFeatureInfo_R(SEXP handle, SEXP field, SEXP features) {
R_API_BEGIN();
SEXP field_char = Rf_protect(Rf_asChar(field));
bst_ulong len_features = Rf_xlength(features);
int res_code;
{
std::vector<void*> dvec(len);
for (R_xlen_t i = 0; i < len; ++i) {
dvec[i] = R_ExternalPtrAddr(VECTOR_ELT(dmats, i));
std::vector<const char*> str_arr(len_features);
for (bst_ulong idx = 0; idx < len_features; idx++) {
str_arr[idx] = CHAR(STRING_ELT(features, idx));
}
res_code = XGBoosterCreate(BeginPtr(dvec), dvec.size(), &handle);
res_code = XGBoosterSetStrFeatureInfo(R_ExternalPtrAddr(handle),
CHAR(field_char),
str_arr.data(),
len_features);
}
CHECK_CALL(res_code);
R_SetExternalPtrAddr(R_handle, handle);
R_RegisterCFinalizerEx(R_handle, _BoosterFinalizer, TRUE);
Rf_unprotect(1);
R_API_END();
return R_NilValue;
}
XGB_DLL SEXP XGBoosterGetStrFeatureInfo_R(SEXP handle, SEXP field) {
R_API_BEGIN();
bst_ulong len;
const char **out_features;
SEXP field_char = Rf_protect(Rf_asChar(field));
CHECK_CALL(XGBoosterGetStrFeatureInfo(R_ExternalPtrAddr(handle),
CHAR(field_char), &len, &out_features));
SEXP out = Rf_protect(Rf_allocVector(STRSXP, len));
for (bst_ulong idx = 0; idx < len; idx++) {
SET_STRING_ELT(out, idx, Rf_mkChar(out_features[idx]));
}
Rf_unprotect(2);
return out;
R_API_END();
return R_NilValue; /* <- should not be reached */
}
XGB_DLL SEXP XGBoosterBoostedRounds_R(SEXP handle) {
SEXP out = Rf_protect(Rf_allocVector(INTSXP, 1));
R_API_BEGIN();
CHECK_CALL(XGBoosterBoostedRounds(R_ExternalPtrAddr(handle), INTEGER(out)));
R_API_END();
Rf_unprotect(1);
return out;
}
/* Note: R's integer class is 32-bit-and-signed only, while xgboost
supports more, so it returns it as a floating point instead */
XGB_DLL SEXP XGBoosterGetNumFeature_R(SEXP handle) {
SEXP out = Rf_protect(Rf_allocVector(REALSXP, 1));
R_API_BEGIN();
bst_ulong res;
CHECK_CALL(XGBoosterGetNumFeature(R_ExternalPtrAddr(handle), &res));
REAL(out)[0] = static_cast<double>(res);
R_API_END();
Rf_unprotect(1);
return out;
}
XGB_DLL SEXP XGBoosterSetParam_R(SEXP handle, SEXP name, SEXP val) {
R_API_BEGIN();
SEXP name_ = PROTECT(Rf_asChar(name));
@ -503,8 +952,8 @@ XGB_DLL SEXP XGBoosterSetParam_R(SEXP handle, SEXP name, SEXP val) {
XGB_DLL SEXP XGBoosterUpdateOneIter_R(SEXP handle, SEXP iter, SEXP dtrain) {
R_API_BEGIN();
CHECK_CALL(XGBoosterUpdateOneIter(R_ExternalPtrAddr(handle),
asInteger(iter),
R_ExternalPtrAddr(dtrain)));
Rf_asInteger(iter),
R_ExternalPtrAddr(dtrain)));
R_API_END();
return R_NilValue;
}
@ -513,20 +962,14 @@ XGB_DLL SEXP XGBoosterTrainOneIter_R(SEXP handle, SEXP dtrain, SEXP iter, SEXP g
R_API_BEGIN();
CHECK_EQ(Rf_xlength(grad), Rf_xlength(hess)) << "gradient and hess must have same length.";
SEXP gdim = getAttrib(grad, R_DimSymbol);
auto n_samples = static_cast<std::size_t>(INTEGER(gdim)[0]);
auto n_targets = static_cast<std::size_t>(INTEGER(gdim)[1]);
SEXP hdim = getAttrib(hess, R_DimSymbol);
CHECK_EQ(INTEGER(hdim)[0], n_samples) << "mismatched size between gradient and hessian";
CHECK_EQ(INTEGER(hdim)[1], n_targets) << "mismatched size between gradient and hessian";
double const *d_grad = REAL(grad);
double const *d_hess = REAL(hess);
int res_code;
{
auto ctx = xgboost::detail::BoosterCtx(R_ExternalPtrAddr(handle));
auto [s_grad, s_hess] = xgboost::detail::MakeGradientInterface(
ctx, d_grad, d_hess, xgboost::linalg::kF, n_samples, n_targets);
const std::string s_grad = Rf_isNull(gdim)?
MakeArrayInterfaceFromRVector(grad) : MakeArrayInterfaceFromRMat(grad);
const std::string s_hess = Rf_isNull(hdim)?
MakeArrayInterfaceFromRVector(hess) : MakeArrayInterfaceFromRMat(hess);
res_code = XGBoosterTrainOneIter(R_ExternalPtrAddr(handle), R_ExternalPtrAddr(dtrain),
asInteger(iter), s_grad.c_str(), s_hess.c_str());
}

100
R-package/src/xgboost_R.h
View File

@ -8,7 +8,9 @@
#define XGBOOST_R_H_ // NOLINT(*)
#include <R.h>
#include <Rinternals.h>
#include <R_ext/Altrep.h>
#include <R_ext/Random.h>
#include <Rmath.h>
@ -53,6 +55,16 @@ XGB_DLL SEXP XGDMatrixCreateFromFile_R(SEXP fname, SEXP silent);
XGB_DLL SEXP XGDMatrixCreateFromMat_R(SEXP mat,
SEXP missing,
SEXP n_threads);
/**
* @brief Create matrix content from a data frame.
* @param data R data.frame object
* @param missing which value to represent missing value
* @param n_threads Number of threads used to construct DMatrix from dense matrix.
* @return created dmatrix
*/
XGB_DLL SEXP XGDMatrixCreateFromDF_R(SEXP df, SEXP missing, SEXP n_threads);
/*!
* \brief create a matrix content from CSC format
* \param indptr pointer to column headers
@ -106,12 +118,20 @@ XGB_DLL SEXP XGDMatrixSaveBinary_R(SEXP handle, SEXP fname, SEXP silent);
XGB_DLL SEXP XGDMatrixSetInfo_R(SEXP handle, SEXP field, SEXP array);
/*!
* \brief get info vector from matrix
* \brief get info vector (float type) from matrix
* \param handle a instance of data matrix
* \param field field name
* \return info vector
*/
XGB_DLL SEXP XGDMatrixGetInfo_R(SEXP handle, SEXP field);
XGB_DLL SEXP XGDMatrixGetFloatInfo_R(SEXP handle, SEXP field);
/*!
* \brief get info vector (uint type) from matrix
* \param handle a instance of data matrix
* \param field field name
* \return info vector
*/
XGB_DLL SEXP XGDMatrixGetUIntInfo_R(SEXP handle, SEXP field);
/*!
* \brief return number of rows
@ -125,19 +145,87 @@ XGB_DLL SEXP XGDMatrixNumRow_R(SEXP handle);
*/
XGB_DLL SEXP XGDMatrixNumCol_R(SEXP handle);
/*!
* \brief Call R C-level function 'duplicate'
* \param obj Object to duplicate
*/
XGB_DLL SEXP XGDuplicate_R(SEXP obj);
/*!
* \brief Equality comparison for two pointers
* \param obj1 R 'externalptr'
* \param obj2 R 'externalptr'
*/
XGB_DLL SEXP XGPointerEqComparison_R(SEXP obj1, SEXP obj2);
/*!
* \brief Register the Altrep class used for the booster
* \param dll DLL info as provided by R_init
*/
XGB_DLL void XGBInitializeAltrepClass_R(DllInfo *dll);
/*!
* \brief return the quantile cuts used for the histogram method
* \param handle an instance of data matrix
* \return A list with entries 'indptr' and 'data'
*/
XGB_DLL SEXP XGDMatrixGetQuantileCut_R(SEXP handle);
/*!
* \brief get the number of non-missing entries in a dmatrix
* \param handle an instance of data matrix
* \return the number of non-missing entries
*/
XGB_DLL SEXP XGDMatrixNumNonMissing_R(SEXP handle);
/*!
* \brief get the data in a dmatrix in CSR format
* \param handle an instance of data matrix
* \return R list with the following entries in this order:
* - 'indptr
* - 'indices
* - 'data'
* - 'ncol'
*/
XGB_DLL SEXP XGDMatrixGetDataAsCSR_R(SEXP handle);
/*!
* \brief create xgboost learner
* \param dmats a list of dmatrix handles that will be cached
*/
XGB_DLL SEXP XGBoosterCreate_R(SEXP dmats);
/*!
* \brief copy information about features from a DMatrix into a Booster
* \param booster R 'externalptr' pointing to a booster object
* \param dmat R 'externalptr' pointing to a DMatrix object
*/
XGB_DLL SEXP XGBoosterCopyInfoFromDMatrix_R(SEXP booster, SEXP dmat);
/*!
* \brief create xgboost learner, saving the pointer into an existing R object
* \param dmats a list of dmatrix handles that will be cached
* \param R_handle a clean R external pointer (not holding any object)
* \brief handle R 'externalptr' holding the booster object
* \param field field name
* \param features features to set for the field
*/
XGB_DLL SEXP XGBoosterCreateInEmptyObj_R(SEXP dmats, SEXP R_handle);
XGB_DLL SEXP XGBoosterSetStrFeatureInfo_R(SEXP handle, SEXP field, SEXP features);
/*!
* \brief handle R 'externalptr' holding the booster object
* \param field field name
*/
XGB_DLL SEXP XGBoosterGetStrFeatureInfo_R(SEXP handle, SEXP field);
/*!
* \brief Get the number of boosted rounds from a model
* \param handle R 'externalptr' holding the booster object
*/
XGB_DLL SEXP XGBoosterBoostedRounds_R(SEXP handle);
/*!
* \brief Get the number of features to which the model was fitted
* \param handle R 'externalptr' holding the booster object
*/
XGB_DLL SEXP XGBoosterGetNumFeature_R(SEXP handle);
/*!
* \brief set parameters

6
R-package/src/xgboost_custom.cc
View File

@ -17,7 +17,11 @@ namespace xgboost {
ConsoleLogger::~ConsoleLogger() {
if (cur_verbosity_ == LogVerbosity::kIgnore ||
cur_verbosity_ <= GlobalVerbosity()) {
dmlc::CustomLogMessage::Log(log_stream_.str());
if (cur_verbosity_ == LogVerbosity::kWarning) {
REprintf("%s\n", log_stream_.str().c_str());
} else {
dmlc::CustomLogMessage::Log(log_stream_.str());
}
}
}
TrackerLogger::~TrackerLogger() {

1
R-package/tests/helper_scripts/install_deps.R
View File

@ -3,7 +3,6 @@
## inconsistent is found.
pkgs <- c(
## CI
"caret",
"pkgbuild",
"roxygen2",
"XML",

418
R-package/tests/testthat/test_basic.R
View File

@ -16,18 +16,19 @@ n_threads <- 1
test_that("train and predict binary classification", {
nrounds <- 2
expect_output(
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = nrounds,
objective = "binary:logistic", eval_metric = "error"
objective = "binary:logistic", eval_metric = "error",
watchlist = list(train = xgb.DMatrix(train$data, label = train$label))
),
"train-error"
)
expect_equal(class(bst), "xgb.Booster")
expect_equal(bst$niter, nrounds)
expect_false(is.null(bst$evaluation_log))
expect_equal(nrow(bst$evaluation_log), nrounds)
expect_lt(bst$evaluation_log[, min(train_error)], 0.03)
expect_equal(xgb.get.num.boosted.rounds(bst), nrounds)
expect_false(is.null(attributes(bst)$evaluation_log))
expect_equal(nrow(attributes(bst)$evaluation_log), nrounds)
expect_lt(attributes(bst)$evaluation_log[, min(train_error)], 0.03)
pred <- predict(bst, test$data)
expect_length(pred, 1611)
@ -35,7 +36,7 @@ test_that("train and predict binary classification", {
pred1 <- predict(bst, train$data, ntreelimit = 1)
expect_length(pred1, 6513)
err_pred1 <- sum((pred1 > 0.5) != train$label) / length(train$label)
err_log <- bst$evaluation_log[1, train_error]
err_log <- attributes(bst)$evaluation_log[1, train_error]
expect_lt(abs(err_pred1 - err_log), 10e-6)
pred2 <- predict(bst, train$data, iterationrange = c(1, 2))
@ -56,7 +57,7 @@ test_that("parameter validation works", {
y <- d[, "x1"] + d[, "x2"]^2 +
ifelse(d[, "x3"] > .5, d[, "x3"]^2, 2^d[, "x3"]) +
rnorm(10)
dtrain <- xgb.DMatrix(data = d, info = list(label = y), nthread = n_threads)
dtrain <- xgb.DMatrix(data = d, label = y, nthread = n_threads)
correct <- function() {
params <- list(
@ -82,7 +83,8 @@ test_that("parameter validation works", {
bar = "foo"
)
output <- capture.output(
xgb.train(params = params, data = dtrain, nrounds = nrounds)
xgb.train(params = params, data = dtrain, nrounds = nrounds),
type = "message"
)
print(output)
}
@ -104,9 +106,8 @@ test_that("dart prediction works", {
rnorm(100)
set.seed(1994)
booster_by_xgboost <- xgboost(
data = d,
label = y,
booster_by_xgboost <- xgb.train(
data = xgb.DMatrix(d, label = y),
max_depth = 2,
booster = "dart",
rate_drop = 0.5,
@ -124,7 +125,7 @@ test_that("dart prediction works", {
expect_false(all(matrix(pred_by_xgboost_0, byrow = TRUE) == matrix(pred_by_xgboost_2, byrow = TRUE)))
set.seed(1994)
dtrain <- xgb.DMatrix(data = d, info = list(label = y), nthread = n_threads)
dtrain <- xgb.DMatrix(data = d, label = y, nthread = n_threads)
booster_by_train <- xgb.train(
params = list(
booster = "dart",
@ -151,16 +152,17 @@ test_that("train and predict softprob", {
lb <- as.numeric(iris$Species) - 1
set.seed(11)
expect_output(
bst <- xgboost(
data = as.matrix(iris[, -5]), label = lb,
bst <- xgb.train(
data = xgb.DMatrix(as.matrix(iris[, -5]), label = lb),
max_depth = 3, eta = 0.5, nthread = n_threads, nrounds = 5,
objective = "multi:softprob", num_class = 3, eval_metric = "merror"
objective = "multi:softprob", num_class = 3, eval_metric = "merror",
watchlist = list(train = xgb.DMatrix(as.matrix(iris[, -5]), label = lb))
),
"train-merror"
)
expect_false(is.null(bst$evaluation_log))
expect_lt(bst$evaluation_log[, min(train_merror)], 0.025)
expect_equal(bst$niter * 3, xgb.ntree(bst))
expect_false(is.null(attributes(bst)$evaluation_log))
expect_lt(attributes(bst)$evaluation_log[, min(train_merror)], 0.025)
expect_equal(xgb.get.num.boosted.rounds(bst) * 3, xgb.ntree(bst))
pred <- predict(bst, as.matrix(iris[, -5]))
expect_length(pred, nrow(iris) * 3)
# row sums add up to total probability of 1:
@ -170,12 +172,12 @@ test_that("train and predict softprob", {
expect_equal(as.numeric(t(mpred)), pred)
pred_labels <- max.col(mpred) - 1
err <- sum(pred_labels != lb) / length(lb)
expect_equal(bst$evaluation_log[5, train_merror], err, tolerance = 5e-6)
expect_equal(attributes(bst)$evaluation_log[5, train_merror], err, tolerance = 5e-6)
# manually calculate error at the 1st iteration:
mpred <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE, ntreelimit = 1)
pred_labels <- max.col(mpred) - 1
err <- sum(pred_labels != lb) / length(lb)
expect_equal(bst$evaluation_log[1, train_merror], err, tolerance = 5e-6)
expect_equal(attributes(bst)$evaluation_log[1, train_merror], err, tolerance = 5e-6)
mpred1 <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE, iterationrange = c(1, 2))
expect_equal(mpred, mpred1)
@ -186,7 +188,7 @@ test_that("train and predict softprob", {
x3 = rnorm(100)
)
y <- sample.int(10, 100, replace = TRUE) - 1
dtrain <- xgb.DMatrix(data = d, info = list(label = y), nthread = n_threads)
dtrain <- xgb.DMatrix(data = d, label = y, nthread = n_threads)
booster <- xgb.train(
params = list(tree_method = "hist", nthread = n_threads),
data = dtrain, nrounds = 4, num_class = 10,
@ -201,39 +203,41 @@ test_that("train and predict softmax", {
lb <- as.numeric(iris$Species) - 1
set.seed(11)
expect_output(
bst <- xgboost(
data = as.matrix(iris[, -5]), label = lb,
bst <- xgb.train(
data = xgb.DMatrix(as.matrix(iris[, -5]), label = lb),
max_depth = 3, eta = 0.5, nthread = n_threads, nrounds = 5,
objective = "multi:softmax", num_class = 3, eval_metric = "merror"
objective = "multi:softmax", num_class = 3, eval_metric = "merror",
watchlist = list(train = xgb.DMatrix(as.matrix(iris[, -5]), label = lb))
),
"train-merror"
)
expect_false(is.null(bst$evaluation_log))
expect_lt(bst$evaluation_log[, min(train_merror)], 0.025)
expect_equal(bst$niter * 3, xgb.ntree(bst))
expect_false(is.null(attributes(bst)$evaluation_log))
expect_lt(attributes(bst)$evaluation_log[, min(train_merror)], 0.025)
expect_equal(xgb.get.num.boosted.rounds(bst) * 3, xgb.ntree(bst))
pred <- predict(bst, as.matrix(iris[, -5]))
expect_length(pred, nrow(iris))
err <- sum(pred != lb) / length(lb)
expect_equal(bst$evaluation_log[5, train_merror], err, tolerance = 5e-6)
expect_equal(attributes(bst)$evaluation_log[5, train_merror], err, tolerance = 5e-6)
})
test_that("train and predict RF", {
set.seed(11)
lb <- train$label
# single iteration
bst <- xgboost(
data = train$data, label = lb, max_depth = 5,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = lb), max_depth = 5,
nthread = n_threads,
nrounds = 1, objective = "binary:logistic", eval_metric = "error",
num_parallel_tree = 20, subsample = 0.6, colsample_bytree = 0.1
num_parallel_tree = 20, subsample = 0.6, colsample_bytree = 0.1,
watchlist = list(train = xgb.DMatrix(train$data, label = lb))
)
expect_equal(bst$niter, 1)
expect_equal(xgb.get.num.boosted.rounds(bst), 1)
expect_equal(xgb.ntree(bst), 20)
pred <- predict(bst, train$data)
pred_err <- sum((pred > 0.5) != lb) / length(lb)
expect_lt(abs(bst$evaluation_log[1, train_error] - pred_err), 10e-6)
expect_lt(abs(attributes(bst)$evaluation_log[1, train_error] - pred_err), 10e-6)
# expect_lt(pred_err, 0.03)
pred <- predict(bst, train$data, ntreelimit = 20)
@ -248,50 +252,53 @@ test_that("train and predict RF with softprob", {
lb <- as.numeric(iris$Species) - 1
nrounds <- 15
set.seed(11)
bst <- xgboost(
data = as.matrix(iris[, -5]), label = lb,
bst <- xgb.train(
data = xgb.DMatrix(as.matrix(iris[, -5]), label = lb),
max_depth = 3, eta = 0.9, nthread = n_threads, nrounds = nrounds,
objective = "multi:softprob", eval_metric = "merror",
num_class = 3, verbose = 0,
num_parallel_tree = 4, subsample = 0.5, colsample_bytree = 0.5
num_parallel_tree = 4, subsample = 0.5, colsample_bytree = 0.5,
watchlist = list(train = xgb.DMatrix(as.matrix(iris[, -5]), label = lb))
)
expect_equal(bst$niter, 15)
expect_equal(xgb.get.num.boosted.rounds(bst), 15)
expect_equal(xgb.ntree(bst), 15 * 3 * 4)
# predict for all iterations:
pred <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE)
expect_equal(dim(pred), c(nrow(iris), 3))
pred_labels <- max.col(pred) - 1
err <- sum(pred_labels != lb) / length(lb)
expect_equal(bst$evaluation_log[nrounds, train_merror], err, tolerance = 5e-6)
expect_equal(attributes(bst)$evaluation_log[nrounds, train_merror], err, tolerance = 5e-6)
# predict for 7 iterations and adjust for 4 parallel trees per iteration
pred <- predict(bst, as.matrix(iris[, -5]), reshape = TRUE, ntreelimit = 7 * 4)
err <- sum((max.col(pred) - 1) != lb) / length(lb)
expect_equal(bst$evaluation_log[7, train_merror], err, tolerance = 5e-6)
expect_equal(attributes(bst)$evaluation_log[7, train_merror], err, tolerance = 5e-6)
})
test_that("use of multiple eval metrics works", {
expect_output(
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
eval_metric = "error", eval_metric = "auc", eval_metric = "logloss"
eval_metric = "error", eval_metric = "auc", eval_metric = "logloss",
watchlist = list(train = xgb.DMatrix(train$data, label = train$label))
),
"train-error.*train-auc.*train-logloss"
)
expect_false(is.null(bst$evaluation_log))
expect_equal(dim(bst$evaluation_log), c(2, 4))
expect_equal(colnames(bst$evaluation_log), c("iter", "train_error", "train_auc", "train_logloss"))
expect_false(is.null(attributes(bst)$evaluation_log))
expect_equal(dim(attributes(bst)$evaluation_log), c(2, 4))
expect_equal(colnames(attributes(bst)$evaluation_log), c("iter", "train_error", "train_auc", "train_logloss"))
expect_output(
bst2 <- xgboost(
data = train$data, label = train$label, max_depth = 2,
bst2 <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
eval_metric = list("error", "auc", "logloss")
eval_metric = list("error", "auc", "logloss"),
watchlist = list(train = xgb.DMatrix(train$data, label = train$label))
),
"train-error.*train-auc.*train-logloss"
)
expect_false(is.null(bst2$evaluation_log))
expect_equal(dim(bst2$evaluation_log), c(2, 4))
expect_equal(colnames(bst2$evaluation_log), c("iter", "train_error", "train_auc", "train_logloss"))
expect_false(is.null(attributes(bst2)$evaluation_log))
expect_equal(dim(attributes(bst2)$evaluation_log), c(2, 4))
expect_equal(colnames(attributes(bst2)$evaluation_log), c("iter", "train_error", "train_auc", "train_logloss"))
})
@ -311,41 +318,25 @@ test_that("training continuation works", {
# continue for two more:
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0, xgb_model = bst1)
if (!windows_flag && !solaris_flag) {
expect_equal(bst$raw, bst2$raw)
expect_equal(xgb.save.raw(bst), xgb.save.raw(bst2))
}
expect_false(is.null(bst2$evaluation_log))
expect_equal(dim(bst2$evaluation_log), c(4, 2))
expect_equal(bst2$evaluation_log, bst$evaluation_log)
expect_false(is.null(attributes(bst2)$evaluation_log))
expect_equal(dim(attributes(bst2)$evaluation_log), c(4, 2))
expect_equal(attributes(bst2)$evaluation_log, attributes(bst)$evaluation_log)
# test continuing from raw model data
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0, xgb_model = bst1$raw)
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0, xgb_model = xgb.save.raw(bst1))
if (!windows_flag && !solaris_flag) {
expect_equal(bst$raw, bst2$raw)
expect_equal(xgb.save.raw(bst), xgb.save.raw(bst2))
}
expect_equal(dim(bst2$evaluation_log), c(2, 2))
expect_equal(dim(attributes(bst2)$evaluation_log), c(2, 2))
# test continuing from a model in file
xgb.save(bst1, "xgboost.json")
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0, xgb_model = "xgboost.json")
fname <- file.path(tempdir(), "xgboost.json")
xgb.save(bst1, fname)
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0, xgb_model = fname)
if (!windows_flag && !solaris_flag) {
expect_equal(bst$raw, bst2$raw)
expect_equal(xgb.save.raw(bst), xgb.save.raw(bst2))
}
expect_equal(dim(bst2$evaluation_log), c(2, 2))
file.remove("xgboost.json")
})
test_that("model serialization works", {
out_path <- "model_serialization"
dtrain <- xgb.DMatrix(train$data, label = train$label, nthread = n_threads)
watchlist <- list(train = dtrain)
param <- list(objective = "binary:logistic", nthread = n_threads)
booster <- xgb.train(param, dtrain, nrounds = 4, watchlist)
raw <- xgb.serialize(booster)
saveRDS(raw, out_path)
raw <- readRDS(out_path)
loaded <- xgb.unserialize(raw)
raw_from_loaded <- xgb.serialize(loaded)
expect_equal(raw, raw_from_loaded)
file.remove(out_path)
expect_equal(dim(attributes(bst2)$evaluation_log), c(2, 2))
})
test_that("xgb.cv works", {
@ -361,7 +352,7 @@ test_that("xgb.cv works", {
expect_is(cv, "xgb.cv.synchronous")
expect_false(is.null(cv$evaluation_log))
expect_lt(cv$evaluation_log[, min(test_error_mean)], 0.03)
expect_lt(cv$evaluation_log[, min(test_error_std)], 0.008)
expect_lt(cv$evaluation_log[, min(test_error_std)], 0.0085)
expect_equal(cv$niter, 2)
expect_false(is.null(cv$folds) && is.list(cv$folds))
expect_length(cv$folds, 5)
@ -391,8 +382,8 @@ test_that("xgb.cv works with stratified folds", {
test_that("train and predict with non-strict classes", {
# standard dense matrix input
train_dense <- as.matrix(train$data)
bst <- xgboost(
data = train_dense, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train_dense, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
verbose = 0
)
@ -402,8 +393,8 @@ test_that("train and predict with non-strict classes", {
class(train_dense) <- "shmatrix"
expect_true(is.matrix(train_dense))
expect_error(
bst <- xgboost(
data = train_dense, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train_dense, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
verbose = 0
),
@ -416,8 +407,8 @@ test_that("train and predict with non-strict classes", {
class(train_dense) <- c("pphmatrix", "shmatrix")
expect_true(is.matrix(train_dense))
expect_error(
bst <- xgboost(
data = train_dense, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train_dense, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
verbose = 0
),
@ -448,8 +439,8 @@ test_that("max_delta_step works", {
# model with restricted max_delta_step
bst2 <- xgb.train(param, dtrain, nrounds, watchlist, verbose = 1, max_delta_step = 1)
# the no-restriction model is expected to have consistently lower loss during the initial iterations
expect_true(all(bst1$evaluation_log$train_logloss < bst2$evaluation_log$train_logloss))
expect_lt(mean(bst1$evaluation_log$train_logloss) / mean(bst2$evaluation_log$train_logloss), 0.8)
expect_true(all(attributes(bst1)$evaluation_log$train_logloss < attributes(bst2)$evaluation_log$train_logloss))
expect_lt(mean(attributes(bst1)$evaluation_log$train_logloss) / mean(attributes(bst2)$evaluation_log$train_logloss), 0.8)
})
test_that("colsample_bytree works", {
@ -480,8 +471,8 @@ test_that("colsample_bytree works", {
})
test_that("Configuration works", {
bst <- xgboost(
data = train$data, label = train$label, max_depth = 2,
bst <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label), max_depth = 2,
eta = 1, nthread = n_threads, nrounds = 2, objective = "binary:logistic",
eval_metric = "error", eval_metric = "auc", eval_metric = "logloss"
)
@ -521,8 +512,8 @@ test_that("strict_shape works", {
y <- as.numeric(iris$Species) - 1
X <- as.matrix(iris[, -5])
bst <- xgboost(
data = X, label = y,
bst <- xgb.train(
data = xgb.DMatrix(X, label = y),
max_depth = 2, nrounds = n_rounds, nthread = n_threads,
objective = "multi:softprob", num_class = 3, eval_metric = "merror"
)
@ -536,8 +527,8 @@ test_that("strict_shape works", {
X <- agaricus.train$data
y <- agaricus.train$label
bst <- xgboost(
data = X, label = y, max_depth = 2, nthread = n_threads,
bst <- xgb.train(
data = xgb.DMatrix(X, label = y), max_depth = 2, nthread = n_threads,
nrounds = n_rounds, objective = "binary:logistic",
eval_metric = "error", eval_metric = "auc", eval_metric = "logloss"
)
@ -555,8 +546,8 @@ test_that("'predict' accepts CSR data", {
x_csc <- as(X[1L, , drop = FALSE], "CsparseMatrix")
x_csr <- as(x_csc, "RsparseMatrix")
x_spv <- as(x_csc, "sparseVector")
bst <- xgboost(
data = X, label = y, objective = "binary:logistic",
bst <- xgb.train(
data = xgb.DMatrix(X, label = y), objective = "binary:logistic",
nrounds = 5L, verbose = FALSE, nthread = n_threads,
)
p_csc <- predict(bst, x_csc)
@ -565,3 +556,234 @@ test_that("'predict' accepts CSR data", {
expect_equal(p_csc, p_csr)
expect_equal(p_csc, p_spv)
})
test_that("Quantile regression accepts multiple quantiles", {
data(mtcars)
y <- mtcars[, 1]
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(data = x, label = y)
model <- xgb.train(
data = dm,
params = list(
objective = "reg:quantileerror",
tree_method = "exact",
quantile_alpha = c(0.05, 0.5, 0.95),
nthread = n_threads
),
nrounds = 15
)
pred <- predict(model, x, reshape = TRUE)
expect_equal(dim(pred)[1], nrow(x))
expect_equal(dim(pred)[2], 3)
expect_true(all(pred[, 1] <= pred[, 3]))
cors <- cor(y, pred)
expect_true(cors[2] > cors[1])
expect_true(cors[2] > cors[3])
expect_true(cors[2] > 0.85)
})
test_that("Can use multi-output labels with built-in objectives", {
data("mtcars")
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
y_mirrored <- cbind(y, -y)
dm <- xgb.DMatrix(x, label = y_mirrored, nthread = n_threads)
model <- xgb.train(
params = list(
tree_method = "hist",
multi_strategy = "multi_output_tree",
objective = "reg:squarederror",
nthread = n_threads
),
data = dm,
nrounds = 5
)
pred <- predict(model, x, reshape = TRUE)
expect_equal(pred[, 1], -pred[, 2])
expect_true(cor(y, pred[, 1]) > 0.9)
expect_true(cor(y, pred[, 2]) < -0.9)
})
test_that("Can use multi-output labels with custom objectives", {
data("mtcars")
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
y_mirrored <- cbind(y, -y)
dm <- xgb.DMatrix(x, label = y_mirrored, nthread = n_threads)
model <- xgb.train(
params = list(
tree_method = "hist",
multi_strategy = "multi_output_tree",
base_score = 0,
objective = function(pred, dtrain) {
y <- getinfo(dtrain, "label")
grad <- pred - y
hess <- rep(1, nrow(grad) * ncol(grad))
hess <- matrix(hess, nrow = nrow(grad))
return(list(grad = grad, hess = hess))
},
nthread = n_threads
),
data = dm,
nrounds = 5
)
pred <- predict(model, x, reshape = TRUE)
expect_equal(pred[, 1], -pred[, 2])
expect_true(cor(y, pred[, 1]) > 0.9)
expect_true(cor(y, pred[, 2]) < -0.9)
})
test_that("Can use ranking objectives with either 'qid' or 'group'", {
set.seed(123)
x <- matrix(rnorm(100 * 10), nrow = 100)
y <- sample(2, size = 100, replace = TRUE) - 1
qid <- c(rep(1, 20), rep(2, 20), rep(3, 60))
gr <- c(20, 20, 60)
dmat_qid <- xgb.DMatrix(x, label = y, qid = qid)
dmat_gr <- xgb.DMatrix(x, label = y, group = gr)
params <- list(tree_method = "hist",
lambdarank_num_pair_per_sample = 8,
objective = "rank:ndcg",
lambdarank_pair_method = "topk",
nthread = n_threads)
set.seed(123)
model_qid <- xgb.train(params, dmat_qid, nrounds = 5)
set.seed(123)
model_gr <- xgb.train(params, dmat_gr, nrounds = 5)
pred_qid <- predict(model_qid, x)
pred_gr <- predict(model_gr, x)
expect_equal(pred_qid, pred_gr)
})
test_that("Coefficients from gblinear have the expected shape and names", {
# Single-column coefficients
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
mm <- model.matrix(~., data = mtcars[, -1])
dm <- xgb.DMatrix(x, label = y, nthread = 1)
model <- xgb.train(
data = dm,
params = list(
booster = "gblinear",
nthread = 1
),
nrounds = 3
)
coefs <- coef(model)
expect_equal(length(coefs), ncol(x) + 1)
expect_equal(names(coefs), c("(Intercept)", colnames(x)))
pred_auto <- predict(model, x)
pred_manual <- as.numeric(mm %*% coefs)
expect_equal(pred_manual, pred_auto, tolerance = 1e-5)
# Multi-column coefficients
data(iris)
y <- as.numeric(iris$Species) - 1
x <- as.matrix(iris[, -5])
dm <- xgb.DMatrix(x, label = y, nthread = 1)
mm <- model.matrix(~., data = iris[, -5])
model <- xgb.train(
data = dm,
params = list(
booster = "gblinear",
objective = "multi:softprob",
num_class = 3,
nthread = 1
),
nrounds = 3
)
coefs <- coef(model)
expect_equal(nrow(coefs), ncol(x) + 1)
expect_equal(ncol(coefs), 3)
expect_equal(row.names(coefs), c("(Intercept)", colnames(x)))
pred_auto <- predict(model, x, outputmargin = TRUE, reshape = TRUE)
pred_manual <- unname(mm %*% coefs)
expect_equal(pred_manual, pred_auto, tolerance = 1e-7)
})
test_that("Deep copies work as expected", {
data(mtcars)
y <- mtcars$mpg
x <- mtcars[, -1]
dm <- xgb.DMatrix(x, label = y, nthread = 1)
model <- xgb.train(
data = dm,
params = list(nthread = 1),
nrounds = 3
)
xgb.attr(model, "my_attr") <- 100
model_shallow_copy <- model
xgb.attr(model_shallow_copy, "my_attr") <- 333
attr_orig <- xgb.attr(model, "my_attr")
attr_shallow <- xgb.attr(model_shallow_copy, "my_attr")
expect_equal(attr_orig, attr_shallow)
model_deep_copy <- xgb.copy.Booster(model)
xgb.attr(model_deep_copy, "my_attr") <- 444
attr_orig <- xgb.attr(model, "my_attr")
attr_deep <- xgb.attr(model_deep_copy, "my_attr")
expect_false(attr_orig == attr_deep)
})
test_that("Pointer comparison works as expected", {
library(xgboost)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
model <- xgb.train(
params = list(nthread = 1),
data = xgb.DMatrix(x, label = y, nthread = 1),
nrounds = 3
)
model_shallow_copy <- model
expect_true(xgb.is.same.Booster(model, model_shallow_copy))
model_deep_copy <- xgb.copy.Booster(model)
expect_false(xgb.is.same.Booster(model, model_deep_copy))
xgb.attr(model_shallow_copy, "my_attr") <- 111
expect_equal(xgb.attr(model, "my_attr"), "111")
expect_null(xgb.attr(model_deep_copy, "my_attr"))
})
test_that("DMatrix field are set to booster when training", {
set.seed(123)
y <- rnorm(100)
x <- matrix(rnorm(100 * 3), nrow = 100)
x[, 2] <- abs(as.integer(x[, 2]))
dm_unnamed <- xgb.DMatrix(x, label = y, nthread = 1)
dm_feature_names <- xgb.DMatrix(x, label = y, feature_names = c("a", "b", "c"), nthread = 1)
dm_feature_types <- xgb.DMatrix(x, label = y)
setinfo(dm_feature_types, "feature_type", c("q", "c", "q"))
dm_both <- xgb.DMatrix(x, label = y, feature_names = c("a", "b", "c"), nthread = 1)
setinfo(dm_both, "feature_type", c("q", "c", "q"))
params <- list(nthread = 1)
model_unnamed <- xgb.train(data = dm_unnamed, params = params, nrounds = 3)
model_feature_names <- xgb.train(data = dm_feature_names, params = params, nrounds = 3)
model_feature_types <- xgb.train(data = dm_feature_types, params = params, nrounds = 3)
model_both <- xgb.train(data = dm_both, params = params, nrounds = 3)
expect_null(getinfo(model_unnamed, "feature_name"))
expect_equal(getinfo(model_feature_names, "feature_name"), c("a", "b", "c"))
expect_null(getinfo(model_feature_types, "feature_name"))
expect_equal(getinfo(model_both, "feature_name"), c("a", "b", "c"))
expect_null(variable.names(model_unnamed))
expect_equal(variable.names(model_feature_names), c("a", "b", "c"))
expect_null(variable.names(model_feature_types))
expect_equal(variable.names(model_both), c("a", "b", "c"))
expect_null(getinfo(model_unnamed, "feature_type"))
expect_null(getinfo(model_feature_names, "feature_type"))
expect_equal(getinfo(model_feature_types, "feature_type"), c("q", "c", "q"))
expect_equal(getinfo(model_both, "feature_type"), c("q", "c", "q"))
})

98
R-package/tests/testthat/test_callbacks.R
View File

@ -57,7 +57,7 @@ test_that("cb.print.evaluation works as expected", {
expect_output(f5(), "\\[7\\]\ttrain-auc:0.900000\ttest-auc:0.800000")
bst_evaluation_err <- c('train-auc' = 0.1, 'test-auc' = 0.2)
expect_output(f1(), "\\[7\\]\ttrain-auc:0.900000\\+0.100000\ttest-auc:0.800000\\+0.200000")
expect_output(f1(), "\\[7\\]\ttrain-auc:0.900000±0.100000\ttest-auc:0.800000±0.200000")
})
test_that("cb.evaluation.log works as expected", {
@ -111,9 +111,9 @@ test_that("can store evaluation_log without printing", {
expect_silent(
bst <- xgb.train(param, dtrain, nrounds = 10, watchlist, eta = 1, verbose = 0)
)
expect_false(is.null(bst$evaluation_log))
expect_false(is.null(bst$evaluation_log$train_error))
expect_lt(bst$evaluation_log[, min(train_error)], 0.2)
expect_false(is.null(attributes(bst)$evaluation_log))
expect_false(is.null(attributes(bst)$evaluation_log$train_error))
expect_lt(attributes(bst)$evaluation_log[, min(train_error)], 0.2)
})
test_that("cb.reset.parameters works as expected", {
@ -121,34 +121,34 @@ test_that("cb.reset.parameters works as expected", {
# fixed eta
set.seed(111)
bst0 <- xgb.train(param, dtrain, nrounds = 2, watchlist, eta = 0.9, verbose = 0)
expect_false(is.null(bst0$evaluation_log))
expect_false(is.null(bst0$evaluation_log$train_error))
expect_false(is.null(attributes(bst0)$evaluation_log))
expect_false(is.null(attributes(bst0)$evaluation_log$train_error))
# same eta but re-set as a vector parameter in the callback
set.seed(111)
my_par <- list(eta = c(0.9, 0.9))
bst1 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
callbacks = list(cb.reset.parameters(my_par)))
expect_false(is.null(bst1$evaluation_log$train_error))
expect_equal(bst0$evaluation_log$train_error,
bst1$evaluation_log$train_error)
expect_false(is.null(attributes(bst1)$evaluation_log$train_error))
expect_equal(attributes(bst0)$evaluation_log$train_error,
attributes(bst1)$evaluation_log$train_error)
# same eta but re-set via a function in the callback
set.seed(111)
my_par <- list(eta = function(itr, itr_end) 0.9)
bst2 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
callbacks = list(cb.reset.parameters(my_par)))
expect_false(is.null(bst2$evaluation_log$train_error))
expect_equal(bst0$evaluation_log$train_error,
bst2$evaluation_log$train_error)
expect_false(is.null(attributes(bst2)$evaluation_log$train_error))
expect_equal(attributes(bst0)$evaluation_log$train_error,
attributes(bst2)$evaluation_log$train_error)
# different eta re-set as a vector parameter in the callback
set.seed(111)
my_par <- list(eta = c(0.6, 0.5))
bst3 <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
callbacks = list(cb.reset.parameters(my_par)))
expect_false(is.null(bst3$evaluation_log$train_error))
expect_false(all(bst0$evaluation_log$train_error == bst3$evaluation_log$train_error))
expect_false(is.null(attributes(bst3)$evaluation_log$train_error))
expect_false(all(attributes(bst0)$evaluation_log$train_error == attributes(bst3)$evaluation_log$train_error))
# resetting multiple parameters at the same time runs with no error
my_par <- list(eta = c(1., 0.5), gamma = c(1, 2), max_depth = c(4, 8))
@ -166,38 +166,39 @@ test_that("cb.reset.parameters works as expected", {
my_par <- list(eta = c(0., 0.))
bstX <- xgb.train(param, dtrain, nrounds = 2, watchlist, verbose = 0,
callbacks = list(cb.reset.parameters(my_par)))
expect_false(is.null(bstX$evaluation_log$train_error))
er <- unique(bstX$evaluation_log$train_error)
expect_false(is.null(attributes(bstX)$evaluation_log$train_error))
er <- unique(attributes(bstX)$evaluation_log$train_error)
expect_length(er, 1)
expect_gt(er, 0.4)
})
test_that("cb.save.model works as expected", {
files <- c('xgboost_01.json', 'xgboost_02.json', 'xgboost.json')
files <- unname(sapply(files, function(f) file.path(tempdir(), f)))
for (f in files) if (file.exists(f)) file.remove(f)
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, eta = 1, verbose = 0,
save_period = 1, save_name = "xgboost_%02d.json")
expect_true(file.exists('xgboost_01.json'))
expect_true(file.exists('xgboost_02.json'))
b1 <- xgb.load('xgboost_01.json')
save_period = 1, save_name = file.path(tempdir(), "xgboost_%02d.json"))
expect_true(file.exists(files[1]))
expect_true(file.exists(files[2]))
b1 <- xgb.load(files[1])
xgb.parameters(b1) <- list(nthread = 2)
expect_equal(xgb.ntree(b1), 1)
b2 <- xgb.load('xgboost_02.json')
expect_equal(xgb.get.num.boosted.rounds(b1), 1)
b2 <- xgb.load(files[2])
xgb.parameters(b2) <- list(nthread = 2)
expect_equal(xgb.ntree(b2), 2)
expect_equal(xgb.get.num.boosted.rounds(b2), 2)
xgb.config(b2) <- xgb.config(bst)
expect_equal(xgb.config(bst), xgb.config(b2))
expect_equal(bst$raw, b2$raw)
expect_equal(xgb.save.raw(bst), xgb.save.raw(b2))
# save_period = 0 saves the last iteration's model
bst <- xgb.train(param, dtrain, nrounds = 2, watchlist, eta = 1, verbose = 0,
save_period = 0, save_name = 'xgboost.json')
expect_true(file.exists('xgboost.json'))
b2 <- xgb.load('xgboost.json')
save_period = 0, save_name = file.path(tempdir(), 'xgboost.json'))
expect_true(file.exists(files[3]))
b2 <- xgb.load(files[3])
xgb.config(b2) <- xgb.config(bst)
expect_equal(bst$raw, b2$raw)
expect_equal(xgb.save.raw(bst), xgb.save.raw(b2))
for (f in files) if (file.exists(f)) file.remove(f)
})
@ -208,14 +209,14 @@ test_that("early stopping xgb.train works", {
bst <- xgb.train(param, dtrain, nrounds = 20, watchlist, eta = 0.3,
early_stopping_rounds = 3, maximize = FALSE)
, "Stopping. Best iteration")
expect_false(is.null(bst$best_iteration))
expect_lt(bst$best_iteration, 19)
expect_equal(bst$best_iteration, bst$best_ntreelimit)
expect_false(is.null(xgb.attr(bst, "best_iteration")))
expect_lt(xgb.attr(bst, "best_iteration"), 19)
expect_equal(xgb.attr(bst, "best_iteration"), xgb.attr(bst, "best_ntreelimit"))
pred <- predict(bst, dtest)
expect_equal(length(pred), 1611)
err_pred <- err(ltest, pred)
err_log <- bst$evaluation_log[bst$best_iteration, test_error]
err_log <- attributes(bst)$evaluation_log[xgb.attr(bst, "best_iteration"), test_error]
expect_equal(err_log, err_pred, tolerance = 5e-6)
set.seed(11)
@ -223,16 +224,15 @@ test_that("early stopping xgb.train works", {
bst0 <- xgb.train(param, dtrain, nrounds = 20, watchlist, eta = 0.3,
early_stopping_rounds = 3, maximize = FALSE, verbose = 0)
)
expect_equal(bst$evaluation_log, bst0$evaluation_log)
expect_equal(attributes(bst)$evaluation_log, attributes(bst0)$evaluation_log)
xgb.save(bst, "model.bin")
loaded <- xgb.load("model.bin")
fname <- file.path(tempdir(), "model.bin")
xgb.save(bst, fname)
loaded <- xgb.load(fname)
expect_false(is.null(loaded$best_iteration))
expect_equal(loaded$best_iteration, bst$best_ntreelimit)
expect_equal(loaded$best_ntreelimit, bst$best_ntreelimit)
file.remove("model.bin")
expect_false(is.null(xgb.attr(loaded, "best_iteration")))
expect_equal(xgb.attr(loaded, "best_iteration"), xgb.attr(bst, "best_ntreelimit"))
expect_equal(xgb.attr(loaded, "best_ntreelimit"), xgb.attr(bst, "best_ntreelimit"))
})
test_that("early stopping using a specific metric works", {
@ -243,14 +243,14 @@ test_that("early stopping using a specific metric works", {
callbacks = list(cb.early.stop(stopping_rounds = 3, maximize = FALSE,
metric_name = 'test_logloss')))
, "Stopping. Best iteration")
expect_false(is.null(bst$best_iteration))
expect_lt(bst$best_iteration, 19)
expect_equal(bst$best_iteration, bst$best_ntreelimit)
expect_false(is.null(xgb.attr(bst, "best_iteration")))
expect_lt(xgb.attr(bst, "best_iteration"), 19)
expect_equal(xgb.attr(bst, "best_iteration"), xgb.attr(bst, "best_ntreelimit"))
pred <- predict(bst, dtest, ntreelimit = bst$best_ntreelimit)
pred <- predict(bst, dtest, ntreelimit = xgb.attr(bst, "best_ntreelimit"))
expect_equal(length(pred), 1611)
logloss_pred <- sum(-ltest * log(pred) - (1 - ltest) * log(1 - pred)) / length(ltest)
logloss_log <- bst$evaluation_log[bst$best_iteration, test_logloss]
logloss_log <- attributes(bst)$evaluation_log[xgb.attr(bst, "best_iteration"), test_logloss]
expect_equal(logloss_log, logloss_pred, tolerance = 1e-5)
})
@ -265,14 +265,14 @@ test_that("early stopping works with titanic", {
dtx <- model.matrix(~ 0 + ., data = titanic[, c("Pclass", "Sex")])
dty <- titanic$Survived
xgboost::xgboost(
data = dtx,
label = dty,
xgboost::xgb.train(
data = xgb.DMatrix(dtx, label = dty),
objective = "binary:logistic",
eval_metric = "auc",
nrounds = 100,
early_stopping_rounds = 3,
nthread = n_threads
nthread = n_threads,
watchlist = list(train = xgb.DMatrix(dtx, label = dty))
)
expect_true(TRUE) # should not crash

8
R-package/tests/testthat/test_custom_objective.R
View File

@ -35,9 +35,9 @@ num_round <- 2
test_that("custom objective works", {
bst <- xgb.train(param, dtrain, num_round, watchlist)
expect_equal(class(bst), "xgb.Booster")
expect_false(is.null(bst$evaluation_log))
expect_false(is.null(bst$evaluation_log$eval_error))
expect_lt(bst$evaluation_log[num_round, eval_error], 0.03)
expect_false(is.null(attributes(bst)$evaluation_log))
expect_false(is.null(attributes(bst)$evaluation_log$eval_error))
expect_lt(attributes(bst)$evaluation_log[num_round, eval_error], 0.03)
})
test_that("custom objective in CV works", {
@ -50,7 +50,7 @@ test_that("custom objective in CV works", {
test_that("custom objective with early stop works", {
bst <- xgb.train(param, dtrain, 10, watchlist)
expect_equal(class(bst), "xgb.Booster")
train_log <- bst$evaluation_log$train_error
train_log <- attributes(bst)$evaluation_log$train_error
expect_true(all(diff(train_log) <= 0))
})

173
R-package/tests/testthat/test_dmatrix.R
View File

@ -67,20 +67,22 @@ test_that("xgb.DMatrix: NA", {
x[1, "x1"] <- NA
m <- xgb.DMatrix(x, nthread = n_threads)
xgb.DMatrix.save(m, "int.dmatrix")
fname_int <- file.path(tempdir(), "int.dmatrix")
xgb.DMatrix.save(m, fname_int)
x <- matrix(as.numeric(x), nrow = n_samples, ncol = 2)
colnames(x) <- c("x1", "x2")
m <- xgb.DMatrix(x, nthread = n_threads)
xgb.DMatrix.save(m, "float.dmatrix")
fname_float <- file.path(tempdir(), "float.dmatrix")
xgb.DMatrix.save(m, fname_float)
iconn <- file("int.dmatrix", "rb")
fconn <- file("float.dmatrix", "rb")
iconn <- file(fname_int, "rb")
fconn <- file(fname_float, "rb")
expect_equal(file.size("int.dmatrix"), file.size("float.dmatrix"))
expect_equal(file.size(fname_int), file.size(fname_float))
bytes <- file.size("int.dmatrix")
bytes <- file.size(fname_int)
idmatrix <- readBin(iconn, "raw", n = bytes)
fdmatrix <- readBin(fconn, "raw", n = bytes)
@ -90,8 +92,8 @@ test_that("xgb.DMatrix: NA", {
close(iconn)
close(fconn)
file.remove("int.dmatrix")
file.remove("float.dmatrix")
file.remove(fname_int)
file.remove(fname_float)
})
test_that("xgb.DMatrix: saving, loading", {
@ -274,17 +276,19 @@ test_that("xgb.DMatrix: Inf as missing", {
x_nan[2, 1] <- NA_real_
m_inf <- xgb.DMatrix(x_inf, nthread = n_threads, missing = Inf)
xgb.DMatrix.save(m_inf, "inf.dmatrix")
fname_inf <- file.path(tempdir(), "inf.dmatrix")
xgb.DMatrix.save(m_inf, fname_inf)
m_nan <- xgb.DMatrix(x_nan, nthread = n_threads, missing = NA_real_)
xgb.DMatrix.save(m_nan, "nan.dmatrix")
fname_nan <- file.path(tempdir(), "nan.dmatrix")
xgb.DMatrix.save(m_nan, fname_nan)
infconn <- file("inf.dmatrix", "rb")
nanconn <- file("nan.dmatrix", "rb")
infconn <- file(fname_inf, "rb")
nanconn <- file(fname_nan, "rb")
expect_equal(file.size("inf.dmatrix"), file.size("nan.dmatrix"))
expect_equal(file.size(fname_inf), file.size(fname_nan))
bytes <- file.size("inf.dmatrix")
bytes <- file.size(fname_inf)
infdmatrix <- readBin(infconn, "raw", n = bytes)
nandmatrix <- readBin(nanconn, "raw", n = bytes)
@ -294,6 +298,143 @@ test_that("xgb.DMatrix: Inf as missing", {
close(infconn)
close(nanconn)
file.remove("inf.dmatrix")
file.remove("nan.dmatrix")
file.remove(fname_inf)
file.remove(fname_nan)
})
test_that("xgb.DMatrix: error on three-dimensional array", {
set.seed(123)
x <- matrix(rnorm(500), nrow = 50)
y <- rnorm(400)
dim(y) <- c(50, 4, 2)
expect_error(xgb.DMatrix(data = x, label = y))
})
test_that("xgb.DMatrix: can get group for both 'qid' and 'group' constructors", {
set.seed(123)
x <- matrix(rnorm(1000), nrow = 100)
group <- c(20, 20, 60)
qid <- c(rep(1, 20), rep(2, 20), rep(3, 60))
gr_mat <- xgb.DMatrix(x, group = group)
qid_mat <- xgb.DMatrix(x, qid = qid)
info_gr <- getinfo(gr_mat, "group")
info_qid <- getinfo(qid_mat, "group")
expect_equal(info_gr, info_qid)
expected_gr <- c(0, 20, 40, 100)
expect_equal(info_gr, expected_gr)
})
test_that("xgb.DMatrix: data.frame", {
df <- data.frame(
a = (1:4) / 10,
num = c(1, NA, 3, 4),
as.int = as.integer(c(1, 2, 3, 4)),
lo = c(TRUE, FALSE, NA, TRUE),
str.fac = c("a", "b", "d", "c"),
as.fac = as.factor(c(3, 5, 8, 11)),
stringsAsFactors = TRUE
)
m <- xgb.DMatrix(df, enable_categorical = TRUE)
expect_equal(colnames(m), colnames(df))
expect_equal(
getinfo(m, "feature_type"), c("float", "float", "int", "i", "c", "c")
)
expect_error(xgb.DMatrix(df))
df <- data.frame(
missing = c("a", "b", "d", NA),
valid = c("a", "b", "d", "c"),
stringsAsFactors = TRUE
)
m <- xgb.DMatrix(df, enable_categorical = TRUE)
expect_equal(getinfo(m, "feature_type"), c("c", "c"))
})
test_that("xgb.DMatrix: can take multi-dimensional 'base_margin'", {
set.seed(123)
x <- matrix(rnorm(100 * 10), nrow = 100)
y <- matrix(rnorm(100 * 2), nrow = 100)
b <- matrix(rnorm(100 * 2), nrow = 100)
model <- xgb.train(
data = xgb.DMatrix(data = x, label = y, nthread = n_threads),
params = list(
objective = "reg:squarederror",
tree_method = "hist",
multi_strategy = "multi_output_tree",
base_score = 0,
nthread = n_threads
),
nround = 1
)
pred_only_x <- predict(model, x, nthread = n_threads, reshape = TRUE)
pred_w_base <- predict(
model,
xgb.DMatrix(data = x, base_margin = b, nthread = n_threads),
nthread = n_threads,
reshape = TRUE
)
expect_equal(pred_only_x, pred_w_base - b, tolerance = 1e-5)
})
test_that("xgb.DMatrix: number of non-missing matches data", {
x <- matrix(1:10, nrow = 5)
dm1 <- xgb.DMatrix(x)
expect_equal(xgb.get.DMatrix.num.non.missing(dm1), 10)
x[2, 2] <- NA
x[4, 1] <- NA
dm2 <- xgb.DMatrix(x)
expect_equal(xgb.get.DMatrix.num.non.missing(dm2), 8)
})
test_that("xgb.DMatrix: retrieving data as CSR", {
data(mtcars)
dm <- xgb.DMatrix(as.matrix(mtcars))
csr <- xgb.get.DMatrix.data(dm)
expect_equal(dim(csr), dim(mtcars))
expect_equal(colnames(csr), colnames(mtcars))
expect_equal(unname(as.matrix(csr)), unname(as.matrix(mtcars)), tolerance = 1e-6)
})
test_that("xgb.DMatrix: quantile cuts look correct", {
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(x, label = y)
model <- xgb.train(
data = dm,
params = list(
tree_method = "hist",
max_bin = 8,
nthread = 1
),
nrounds = 3
)
qcut_list <- xgb.get.DMatrix.qcut(dm, "list")
qcut_arrays <- xgb.get.DMatrix.qcut(dm, "arrays")
expect_equal(length(qcut_arrays), 2)
expect_equal(names(qcut_arrays), c("indptr", "data"))
expect_equal(length(qcut_arrays$indptr), ncol(x) + 1)
expect_true(min(diff(qcut_arrays$indptr)) > 0)
col_min <- apply(x, 2, min)
col_max <- apply(x, 2, max)
expect_equal(length(qcut_list), ncol(x))
expect_equal(names(qcut_list), colnames(x))
lapply(
seq(1, ncol(x)),
function(col) {
cuts <- qcut_list[[col]]
expect_true(min(diff(cuts)) > 0)
expect_true(col_min[col] > cuts[1])
expect_true(col_max[col] < cuts[length(cuts)])
expect_true(length(cuts) <= 9)
}
)
})

4
R-package/tests/testthat/test_gc_safety.R
View File

@ -6,8 +6,8 @@ test_that("train and prediction when gctorture is on", {
train <- agaricus.train
test <- agaricus.test
gctorture(TRUE)
bst <- xgboost(data = train$data, label = train$label, max.depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
bst <- xgb.train(data = xgb.DMatrix(train$data, label = train$label), max.depth = 2,
eta = 1, nthread = 2, nrounds = 2, objective = "binary:logistic")
pred <- predict(bst, test$data)
gctorture(FALSE)
expect_length(pred, length(test$label))

16
R-package/tests/testthat/test_glm.R
View File

@ -24,28 +24,28 @@ test_that("gblinear works", {
bst <- xgb.train(param, dtrain, n, watchlist, verbose = VERB, feature_selector = 'shuffle')
ypred <- predict(bst, dtest)
expect_equal(length(getinfo(dtest, 'label')), 1611)
expect_lt(bst$evaluation_log$eval_error[n], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[n], ERR_UL)
bst <- xgb.train(param, dtrain, n, watchlist, verbose = VERB, feature_selector = 'cyclic',
callbacks = list(cb.gblinear.history()))
expect_lt(bst$evaluation_log$eval_error[n], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[n], ERR_UL)
h <- xgb.gblinear.history(bst)
expect_equal(dim(h), c(n, ncol(dtrain) + 1))
expect_is(h, "matrix")
param$updater <- 'coord_descent'
bst <- xgb.train(param, dtrain, n, watchlist, verbose = VERB, feature_selector = 'cyclic')
expect_lt(bst$evaluation_log$eval_error[n], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[n], ERR_UL)
bst <- xgb.train(param, dtrain, n, watchlist, verbose = VERB, feature_selector = 'shuffle')
expect_lt(bst$evaluation_log$eval_error[n], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[n], ERR_UL)
bst <- xgb.train(param, dtrain, 2, watchlist, verbose = VERB, feature_selector = 'greedy')
expect_lt(bst$evaluation_log$eval_error[2], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[2], ERR_UL)
bst <- xgb.train(param, dtrain, n, watchlist, verbose = VERB, feature_selector = 'thrifty',
top_k = 50, callbacks = list(cb.gblinear.history(sparse = TRUE)))
expect_lt(bst$evaluation_log$eval_error[n], ERR_UL)
expect_lt(attributes(bst)$evaluation_log$eval_error[n], ERR_UL)
h <- xgb.gblinear.history(bst)
expect_equal(dim(h), c(n, ncol(dtrain) + 1))
expect_s4_class(h, "dgCMatrix")
@ -72,10 +72,10 @@ test_that("gblinear early stopping works", {
booster <- xgb.train(
param, dtrain, n, list(eval = dtest, train = dtrain), early_stopping_rounds = es_round
)
expect_equal(booster$best_iteration, 5)
expect_equal(xgb.attr(booster, "best_iteration"), 5)
predt_es <- predict(booster, dtrain)
n <- booster$best_iteration + es_round
n <- xgb.attr(booster, "best_iteration") + es_round
booster <- xgb.train(
param, dtrain, n, list(eval = dtest, train = dtrain), early_stopping_rounds = es_round
)

111
R-package/tests/testthat/test_helpers.R
View File

@ -25,15 +25,15 @@ if (isTRUE(VCD_AVAILABLE)) {
label <- df[, ifelse(Improved == "Marked", 1, 0)]
# binary
bst.Tree <- xgboost(data = sparse_matrix, label = label, max_depth = 9,
eta = 1, nthread = 2, nrounds = nrounds, verbose = 0,
objective = "binary:logistic", booster = "gbtree",
base_score = 0.5)
bst.Tree <- xgb.train(data = xgb.DMatrix(sparse_matrix, label = label), max_depth = 9,
eta = 1, nthread = 2, nrounds = nrounds, verbose = 0,
objective = "binary:logistic", booster = "gbtree",
base_score = 0.5)
bst.GLM <- xgboost(data = sparse_matrix, label = label,
eta = 1, nthread = 1, nrounds = nrounds, verbose = 0,
objective = "binary:logistic", booster = "gblinear",
base_score = 0.5)
bst.GLM <- xgb.train(data = xgb.DMatrix(sparse_matrix, label = label),
eta = 1, nthread = 1, nrounds = nrounds, verbose = 0,
objective = "binary:logistic", booster = "gblinear",
base_score = 0.5)
feature.names <- colnames(sparse_matrix)
}
@ -41,14 +41,17 @@ if (isTRUE(VCD_AVAILABLE)) {
# multiclass
mlabel <- as.numeric(iris$Species) - 1
nclass <- 3
mbst.Tree <- xgboost(data = as.matrix(iris[, -5]), label = mlabel, verbose = 0,
max_depth = 3, eta = 0.5, nthread = 2, nrounds = nrounds,
objective = "multi:softprob", num_class = nclass, base_score = 0)
mbst.Tree <- xgb.train(data = xgb.DMatrix(as.matrix(iris[, -5]), label = mlabel), verbose = 0,
max_depth = 3, eta = 0.5, nthread = 2, nrounds = nrounds,
objective = "multi:softprob", num_class = nclass, base_score = 0)
mbst.GLM <- xgboost(data = as.matrix(iris[, -5]), label = mlabel, verbose = 0,
booster = "gblinear", eta = 0.1, nthread = 1, nrounds = nrounds,
objective = "multi:softprob", num_class = nclass, base_score = 0)
mbst.GLM <- xgb.train(data = xgb.DMatrix(as.matrix(iris[, -5]), label = mlabel), verbose = 0,
booster = "gblinear", eta = 0.1, nthread = 1, nrounds = nrounds,
objective = "multi:softprob", num_class = nclass, base_score = 0)
# without feature names
bst.Tree.unnamed <- xgb.copy.Booster(bst.Tree)
setinfo(bst.Tree.unnamed, "feature_name", NULL)
test_that("xgb.dump works", {
.skip_if_vcd_not_available()
@ -71,8 +74,9 @@ test_that("xgb.dump works for gblinear", {
expect_length(xgb.dump(bst.GLM), 14)
# also make sure that it works properly for a sparse model where some coefficients
# are 0 from setting large L1 regularization:
bst.GLM.sp <- xgboost(data = sparse_matrix, label = label, eta = 1, nthread = 2, nrounds = 1,
alpha = 2, objective = "binary:logistic", booster = "gblinear")
bst.GLM.sp <- xgb.train(data = xgb.DMatrix(sparse_matrix, label = label), eta = 1,
nthread = 2, nrounds = 1,
alpha = 2, objective = "binary:logistic", booster = "gblinear")
d.sp <- xgb.dump(bst.GLM.sp)
expect_length(d.sp, 14)
expect_gt(sum(d.sp == "0"), 0)
@ -168,7 +172,7 @@ test_that("SHAPs sum to predictions, with or without DART", {
nrounds <- 30
for (booster in list("gbtree", "dart")) {
fit <- xgboost(
fit <- xgb.train(
params = c(
list(
nthread = 2,
@ -177,8 +181,7 @@ test_that("SHAPs sum to predictions, with or without DART", {
eval_metric = "rmse"),
if (booster == "dart")
list(rate_drop = .01, one_drop = TRUE)),
data = d,
label = y,
data = xgb.DMatrix(d, label = y),
nrounds = nrounds)
pr <- function(...) {
@ -204,7 +207,7 @@ test_that("xgb-attribute functionality", {
list.ch <- list.val[order(names(list.val))]
list.ch <- lapply(list.ch, as.character)
# note: iter is 0-index in xgb attributes
list.default <- list(niter = as.character(nrounds - 1))
list.default <- list()
list.ch <- c(list.ch, list.default)
# proper input:
expect_error(xgb.attr(bst.Tree, NULL))
@ -212,24 +215,25 @@ test_that("xgb-attribute functionality", {
# set & get:
expect_null(xgb.attr(bst.Tree, "asdf"))
expect_equal(xgb.attributes(bst.Tree), list.default)
xgb.attr(bst.Tree, "my_attr") <- val
expect_equal(xgb.attr(bst.Tree, "my_attr"), val)
xgb.attributes(bst.Tree) <- list.val
expect_equal(xgb.attributes(bst.Tree), list.ch)
bst.Tree.copy <- xgb.copy.Booster(bst.Tree)
xgb.attr(bst.Tree.copy, "my_attr") <- val
expect_equal(xgb.attr(bst.Tree.copy, "my_attr"), val)
xgb.attributes(bst.Tree.copy) <- list.val
expect_equal(xgb.attributes(bst.Tree.copy), list.ch)
# serializing:
xgb.save(bst.Tree, 'xgb.model')
bst <- xgb.load('xgb.model')
if (file.exists('xgb.model')) file.remove('xgb.model')
fname <- file.path(tempdir(), "xgb.ubj")
xgb.save(bst.Tree.copy, fname)
bst <- xgb.load(fname)
expect_equal(xgb.attr(bst, "my_attr"), val)
expect_equal(xgb.attributes(bst), list.ch)
# deletion:
xgb.attr(bst, "my_attr") <- NULL
expect_null(xgb.attr(bst, "my_attr"))
expect_equal(xgb.attributes(bst), list.ch[c("a", "b", "niter")])
expect_equal(xgb.attributes(bst), list.ch[c("a", "b")])
xgb.attributes(bst) <- list(a = NULL, b = NULL)
expect_equal(xgb.attributes(bst), list.default)
xgb.attributes(bst) <- list(niter = NULL)
expect_null(xgb.attributes(bst))
expect_equal(xgb.attributes(bst), list())
})
if (grepl('Windows', Sys.info()[['sysname']], fixed = TRUE) ||
@ -256,27 +260,23 @@ if (grepl('Windows', Sys.info()[['sysname']], fixed = TRUE) ||
test_that("xgb.Booster serializing as R object works", {
.skip_if_vcd_not_available()
saveRDS(bst.Tree, 'xgb.model.rds')
bst <- readRDS('xgb.model.rds')
fname_rds <- file.path(tempdir(), "xgb.model.rds")
saveRDS(bst.Tree, fname_rds)
bst <- readRDS(fname_rds)
dtrain <- xgb.DMatrix(sparse_matrix, label = label, nthread = 2)
expect_equal(predict(bst.Tree, dtrain), predict(bst, dtrain), tolerance = float_tolerance)
expect_equal(xgb.dump(bst.Tree), xgb.dump(bst))
xgb.save(bst, 'xgb.model')
if (file.exists('xgb.model')) file.remove('xgb.model')
bst <- readRDS('xgb.model.rds')
if (file.exists('xgb.model.rds')) file.remove('xgb.model.rds')
nil_ptr <- new("externalptr")
class(nil_ptr) <- "xgb.Booster.handle"
expect_true(identical(bst$handle, nil_ptr))
bst <- xgb.Booster.complete(bst)
expect_true(!identical(bst$handle, nil_ptr))
fname_bin <- file.path(tempdir(), "xgb.model")
xgb.save(bst, fname_bin)
bst <- readRDS(fname_rds)
expect_equal(predict(bst.Tree, dtrain), predict(bst, dtrain), tolerance = float_tolerance)
})
test_that("xgb.model.dt.tree works with and without feature names", {
.skip_if_vcd_not_available()
names.dt.trees <- c("Tree", "Node", "ID", "Feature", "Split", "Yes", "No", "Missing", "Quality", "Cover")
dt.tree <- xgb.model.dt.tree(feature_names = feature.names, model = bst.Tree)
names.dt.trees <- c("Tree", "Node", "ID", "Feature", "Split", "Yes", "No", "Missing", "Gain", "Cover")
dt.tree <- xgb.model.dt.tree(model = bst.Tree)
expect_equal(names.dt.trees, names(dt.tree))
if (!flag_32bit)
expect_equal(dim(dt.tree), c(188, 10))
@ -286,9 +286,7 @@ test_that("xgb.model.dt.tree works with and without feature names", {
expect_equal(dt.tree, dt.tree.0)
# when model contains no feature names:
bst.Tree.x <- bst.Tree
bst.Tree.x$feature_names <- NULL
dt.tree.x <- xgb.model.dt.tree(model = bst.Tree.x)
dt.tree.x <- xgb.model.dt.tree(model = bst.Tree.unnamed)
expect_output(str(dt.tree.x), 'Feature.*\\"3\\"')
expect_equal(dt.tree[, -4, with = FALSE], dt.tree.x[, -4, with = FALSE])
@ -316,9 +314,7 @@ test_that("xgb.importance works with and without feature names", {
expect_equal(importance.Tree, importance.Tree.0, tolerance = float_tolerance)
# when model contains no feature names:
bst.Tree.x <- bst.Tree
bst.Tree.x$feature_names <- NULL
importance.Tree.x <- xgb.importance(model = bst.Tree)
importance.Tree.x <- xgb.importance(model = bst.Tree.unnamed)
expect_equal(importance.Tree[, -1, with = FALSE], importance.Tree.x[, -1, with = FALSE],
tolerance = float_tolerance)
@ -334,14 +330,14 @@ test_that("xgb.importance works with and without feature names", {
importance <- xgb.importance(feature_names = feature.names, model = bst.Tree, trees = trees)
importance_from_dump <- function() {
model_text_dump <- xgb.dump(model = bst.Tree, with_stats = TRUE, trees = trees)
model_text_dump <- xgb.dump(model = bst.Tree.unnamed, with_stats = TRUE, trees = trees)
imp <- xgb.model.dt.tree(
feature_names = feature.names,
text = model_text_dump,
trees = trees
)[
Feature != "Leaf", .(
Gain = sum(Quality),
Gain = sum(Gain),
Cover = sum(Cover),
Frequency = .N
),
@ -360,9 +356,8 @@ test_that("xgb.importance works with and without feature names", {
expect_equal(importance_from_dump(), importance, tolerance = 1e-6)
## decision stump
m <- xgboost::xgboost(
data = as.matrix(data.frame(x = c(0, 1))),
label = c(1, 2),
m <- xgboost::xgb.train(
data = xgb.DMatrix(as.matrix(data.frame(x = c(0, 1))), label = c(1, 2)),
nrounds = 1,
base_score = 0.5,
nthread = 2
@ -393,9 +388,9 @@ test_that("xgb.importance works with GLM model", {
test_that("xgb.model.dt.tree and xgb.importance work with a single split model", {
.skip_if_vcd_not_available()
bst1 <- xgboost(data = sparse_matrix, label = label, max_depth = 1,
eta = 1, nthread = 2, nrounds = 1, verbose = 0,
objective = "binary:logistic")
bst1 <- xgb.train(data = xgb.DMatrix(sparse_matrix, label = label), max_depth = 1,
eta = 1, nthread = 2, nrounds = 1, verbose = 0,
objective = "binary:logistic")
expect_error(dt <- xgb.model.dt.tree(model = bst1), regexp = NA) # no error
expect_equal(nrow(dt), 3)
expect_error(imp <- xgb.importance(model = bst1), regexp = NA) # no error
@ -415,13 +410,13 @@ test_that("xgb.plot.importance de-duplicates features", {
test_that("xgb.plot.tree works with and without feature names", {
.skip_if_vcd_not_available()
expect_silent(xgb.plot.tree(feature_names = feature.names, model = bst.Tree))
expect_silent(xgb.plot.tree(feature_names = feature.names, model = bst.Tree.unnamed))
expect_silent(xgb.plot.tree(model = bst.Tree))
})
test_that("xgb.plot.multi.trees works with and without feature names", {
.skip_if_vcd_not_available()
xgb.plot.multi.trees(model = bst.Tree, feature_names = feature.names, features_keep = 3)
xgb.plot.multi.trees(model = bst.Tree.unnamed, feature_names = feature.names, features_keep = 3)
xgb.plot.multi.trees(model = bst.Tree, features_keep = 3)
})

8
R-package/tests/testthat/test_interaction_constraints.R
View File

@ -13,9 +13,9 @@ train <- matrix(c(x1, x2, x3), ncol = 3)
test_that("interaction constraints for regression", {
# Fit a model that only allows interaction between x1 and x2
bst <- xgboost(data = train, label = y, max_depth = 3,
eta = 0.1, nthread = 2, nrounds = 100, verbose = 0,
interaction_constraints = list(c(0, 1)))
bst <- xgb.train(data = xgb.DMatrix(train, label = y), max_depth = 3,
eta = 0.1, nthread = 2, nrounds = 100, verbose = 0,
interaction_constraints = list(c(0, 1)))
# Set all observations to have the same x3 values then increment
# by the same amount
@ -47,7 +47,7 @@ test_that("interaction constraints scientific representation", {
d <- matrix(rexp(rows, rate = .1), nrow = rows, ncol = cols)
y <- rnorm(rows)
dtrain <- xgb.DMatrix(data = d, info = list(label = y), nthread = n_threads)
dtrain <- xgb.DMatrix(data = d, label = y, nthread = n_threads)
inc <- list(c(seq.int(from = 0, to = cols, by = 1)))
with_inc <- xgb.train(

10
R-package/tests/testthat/test_interactions.R
View File

@ -98,15 +98,14 @@ test_that("SHAP contribution values are not NAN", {
ivs <- c("x1", "x2")
fit <- xgboost(
fit <- xgb.train(
verbose = 0,
params = list(
objective = "reg:squarederror",
eval_metric = "rmse",
nthread = n_threads
),
data = as.matrix(subset(d, fold == 2)[, ivs]),
label = subset(d, fold == 2)$y,
data = xgb.DMatrix(as.matrix(subset(d, fold == 2)[, ivs]), label = subset(d, fold == 2)$y),
nrounds = 3
)
@ -169,9 +168,8 @@ test_that("multiclass feature interactions work", {
test_that("SHAP single sample works", {
train <- agaricus.train
test <- agaricus.test
booster <- xgboost(
data = train$data,
label = train$label,
booster <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label),
max_depth = 2,
nrounds = 4,
objective = "binary:logistic",

51
R-package/tests/testthat/test_io.R
View File

@ -7,8 +7,8 @@ test <- agaricus.test
test_that("load/save raw works", {
nrounds <- 8
booster <- xgboost(
data = train$data, label = train$label,
booster <- xgb.train(
data = xgb.DMatrix(train$data, label = train$label),
nrounds = nrounds, objective = "binary:logistic",
nthread = 2
)
@ -17,8 +17,8 @@ test_that("load/save raw works", {
ubj_bytes <- xgb.save.raw(booster, raw_format = "ubj")
old_bytes <- xgb.save.raw(booster, raw_format = "deprecated")
from_json <- xgb.load.raw(json_bytes, as_booster = TRUE)
from_ubj <- xgb.load.raw(ubj_bytes, as_booster = TRUE)
from_json <- xgb.load.raw(json_bytes)
from_ubj <- xgb.load.raw(ubj_bytes)
json2old <- xgb.save.raw(from_json, raw_format = "deprecated")
ubj2old <- xgb.save.raw(from_ubj, raw_format = "deprecated")
@ -26,3 +26,46 @@ test_that("load/save raw works", {
expect_equal(json2old, ubj2old)
expect_equal(json2old, old_bytes)
})
test_that("saveRDS preserves C and R attributes", {
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(x, label = y, nthread = 1)
model <- xgb.train(
data = dm,
params = list(nthread = 1, max_depth = 2),
nrounds = 5
)
attributes(model)$my_attr <- "qwerty"
xgb.attr(model, "c_attr") <- "asdf"
fname <- file.path(tempdir(), "xgb_model.Rds")
saveRDS(model, fname)
model_new <- readRDS(fname)
expect_equal(attributes(model_new)$my_attr, attributes(model)$my_attr)
expect_equal(xgb.attr(model, "c_attr"), xgb.attr(model_new, "c_attr"))
})
test_that("R serializers keep C config", {
data(mtcars)
y <- mtcars$mpg
x <- as.matrix(mtcars[, -1])
dm <- xgb.DMatrix(x, label = y, nthread = 1)
model <- xgb.train(
data = dm,
params = list(
tree_method = "approx",
nthread = 1,
max_depth = 2
),
nrounds = 3
)
model_new <- unserialize(serialize(model, NULL))
expect_equal(
xgb.config(model)$learner$gradient_booster$gbtree_train_param$tree_method,
xgb.config(model_new)$learner$gradient_booster$gbtree_train_param$tree_method
)
expect_equal(variable.names(model), variable.names(model_new))
})

10
R-package/tests/testthat/test_model_compatibility.R
View File

@ -23,11 +23,7 @@ get_num_tree <- function(booster) {
}
run_booster_check <- function(booster, name) {
# If given a handle, we need to call xgb.Booster.complete() prior to using xgb.config().
if (inherits(booster, "xgb.Booster") && xgboost:::is.null.handle(booster$handle)) {
booster <- xgb.Booster.complete(booster)
}
config <- jsonlite::fromJSON(xgb.config(booster))
config <- xgb.config(booster)
run_model_param_check(config)
if (name == 'cls') {
testthat::expect_equal(get_num_tree(booster),
@ -76,6 +72,10 @@ test_that("Models from previous versions of XGBoost can be loaded", {
name <- m[3]
is_rds <- endsWith(model_file, '.rds')
is_json <- endsWith(model_file, '.json')
# TODO: update this test for new RDS format
if (is_rds) {
return(NULL)
}
# Expect an R warning when a model is loaded from RDS and it was generated by version < 1.1.x
if (is_rds && compareVersion(model_xgb_ver, '1.1.1.1') < 0) {
booster <- readRDS(model_file)

6
R-package/tests/testthat/test_monotone.R
View File

@ -7,9 +7,9 @@ train <- matrix(x, ncol = 1)
test_that("monotone constraints for regression", {
bst <- xgboost(data = train, label = y, max_depth = 2,
eta = 0.1, nthread = 2, nrounds = 100, verbose = 0,
monotone_constraints = -1)
bst <- xgb.train(data = xgb.DMatrix(train, label = y), max_depth = 2,
eta = 0.1, nthread = 2, nrounds = 100, verbose = 0,
monotone_constraints = -1)
pred <- predict(bst, train)

Some files were not shown because too many files have changed in this diff Show More