Merge branch 'master' of https://github.com/tqchen/xgboost

R-package/NAMESPACE

@@ -8,6 +8,7 @@ export(xgb.dump)
 export(xgb.load)
 export(xgb.save)
 export(xgb.train)
+export(xgb.cv)
 export(xgboost)
 exportMethods(predict)
 import(methods)

R-package/R/utils.R

@@ -103,6 +103,10 @@ xgb.get.DMatrix <- function(data, label = NULL) {
   }
   return (dtrain)
 }
+xgb.numrow <- function(dmat) {
+  nrow <- .Call("XGDMatrixNumRow_R", dmat, PACKAGE="xgboost")
+  return(nrow)
+}
 # iteratively update booster with customized statistics
 xgb.iter.boost <- function(booster, dtrain, gpair) {
   if (class(booster) != "xgb.Booster") {
@@ -174,23 +178,51 @@ xgb.iter.eval <- function(booster, watchlist, iter, feval = NULL) {
     }
   } else {
     msg <- ""
-  }  
+  }
   return(msg)
 } 
 #------------------------------------------
 # helper functions for cross validation
 #
-xgb.cv.mknfold <- function(dall, nfold, param, metrics=list(), fpreproc = NULL) {  
+xgb.cv.mknfold <- function(dall, nfold, param) {
   randidx <- sample(1 : xgb.numrow(dall))
   kstep <- length(randidx) / nfold
   idset <- list()
   for (i in 1:nfold) {
-    idset = append(idset, randidx[ ((i-1) * kstep + 1) : min(i * kstep, length(randidx)) ])
+    idset[[i]] <- randidx[ ((i-1) * kstep + 1) : min(i * kstep, length(randidx)) ]
   }
   ret <- list()
   for (k in 1:nfold) {
-    
+    dtest <- slice(dall, idset[[k]])
+    didx = c()
+    for (i in 1:nfold) {
+      if (i != k) {
+        didx <- append(didx, idset[[i]])
+      }
+    }
+    dtrain <- slice(dall, didx)
+    bst <- xgb.Booster(param, list(dtrain, dtest))
+    watchlist = list(train=dtrain, test=dtest)
+    ret[[k]] <- list(dtrain=dtrain, booster=bst, watchlist=watchlist)
   }
-  
+  return (ret)
+}
+xgb.cv.aggcv <- function(res, showsd = TRUE) {
+  header <- res[[1]]
+  ret <- header[1]
+  for (i in 2:length(header)) {
+    kv <- strsplit(header[i], ":")[[1]]
+    ret <- paste(ret, "\t", kv[1], ":", sep="")
+    stats <- c()
+    stats[1] <- as.numeric(kv[2])    
+    for (j in 2:length(res)) {
+      tkv <- strsplit(res[[j]][i], ":")[[1]]
+      stats[j] <- as.numeric(tkv[2])
+    }
+    ret <- paste(ret, sprintf("%f", mean(stats)), sep="")
+    if (showsd) {
+      ret <- paste(ret, sprintf("+%f", sd(stats)), sep="")
+    }
+  }
+  return (ret)
 }
-

R-package/R/xgb.cv.R

@@ -18,6 +18,9 @@
 #'   further details. See also inst/examples/demo.R for walkthrough example in R.
 #' @param data takes an \code{xgb.DMatrix} as the input.
 #' @param nrounds the max number of iterations
+#' @param nfold number of folds used
+#' @param label option field, when data is Matrix
+#' @param showd boolean, whether show standard deviation of cross validation
 #' @param metrics, list of evaluation metrics to be used in corss validation,
 #'   when it is not specified, the evaluation metric is chosen according to objective function.
 #'   Possible options are:
@@ -28,7 +31,6 @@
 #'   \item \code{auc} Area under curve
 #'   \item \code{merror} Exact matching error, used to evaluate multi-class classification
 #' }
-#'
 #' @param obj customized objective function. Returns gradient and second order 
 #'   gradient with given prediction and dtrain, 
 #' @param feval custimized evaluation function. Returns 
@@ -46,12 +48,33 @@
 #'
 #' @export
 #'
-xgb.cv <- function(params=list(), data, nrounds, metrics=list(), label = NULL,
-                   obj = NULL, feval = NULL, ...) {
+xgb.cv <- function(params=list(), data, nrounds, nfold, label = NULL,
+                   showsd = TRUE, metrics=list(), obj = NULL, feval = NULL, ...) {
   if (typeof(params) != "list") {
     stop("xgb.cv: first argument params must be list")
   }
+  if (nfold <= 1) {
+    stop("nfold must be bigger than 1")
+  }
   dtrain <- xgb.get.DMatrix(data, label)
-  params = append(params, list(...))
-  
+  params <- append(params, list(...))
+  params <- append(params, list(silent=1))
+  for (mc in metrics) {
+    params <- append(params, list("eval_metric"=mc))
+  }    
+
+  folds <- xgb.cv.mknfold(dtrain, nfold, params)
+  history <- list()
+  for (i in 1:nrounds) {
+    msg <- list()
+    for (k in 1:nfold) {
+      fd <- folds[[k]]
+      succ <- xgb.iter.update(fd$booster, fd$dtrain, i - 1, obj)      
+      msg[[k]] <- strsplit(xgb.iter.eval(fd$booster, fd$watchlist, i - 1, feval), "\t")[[1]]
+    }
+    ret <- xgb.cv.aggcv(msg, showsd)
+    history <- append(history, ret)
+    cat(paste(ret, "\n", sep=""))
+  }
+  return (history)
 }

R-package/inst/examples/cross_validation.R

@@ -0,0 +1,10 @@
+require(xgboost)
+require(methods)
+# Directly read in local file
+dtrain <- xgb.DMatrix("agaricus.txt.train")
+
+history <- xgb.cv( data = dtrain, nround=3, nfold = 5, metrics=list("rmse","auc"),
+                  "max_depth"=3, "eta"=1,
+                  "objective"="binary:logistic")
+
+

R-package/src/xgboost_R.cpp

@@ -174,6 +174,10 @@ extern "C" {
     _WrapperEnd();
     return ret;
   }
+  SEXP XGDMatrixNumRow_R(SEXP handle) {
+    bst_ulong nrow = XGDMatrixNumRow(R_ExternalPtrAddr(handle));
+    return ScalarInteger(static_cast<int>(nrow));
+  }
   // functions related to booster
   void _BoosterFinalizer(SEXP ext) {    
     if (R_ExternalPtrAddr(ext) == NULL) return;

R-package/src/xgboost_R.h

@@ -65,6 +65,11 @@ extern "C" {
    * \return info vector
    */  
   SEXP XGDMatrixGetInfo_R(SEXP handle, SEXP field);
+  /*!
+   * \brief return number of rows
+   * \param handle a instance of data matrix
+   */
+  SEXP XGDMatrixNumRow_R(SEXP handle);
   /*! 
    * \brief create xgboost learner 
    * \param dmats a list of dmatrix handles that will be cached