[R] Implement feature weights. (#7660)

R-package/R/xgb.DMatrix.R

@@ -287,6 +287,13 @@ setinfo.xgb.DMatrix <- function(object, name, info, ...) {
     .Call(XGDMatrixSetInfo_R, object, name, as.integer(info))
     return(TRUE)
   }
+  if (name == "feature_weights") {
+    if (length(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))
+    return(TRUE)
+  }
   stop("setinfo: unknown info name ", name)
   return(FALSE)
 }

R-package/tests/testthat/test_feature_weights.R

@@ -0,0 +1,27 @@
+library(xgboost)
+
+context("feature weights")
+
+test_that("training with feature weights works", {
+  nrows <- 1000
+  ncols <- 9
+  set.seed(2022)
+  x <- matrix(rnorm(nrows * ncols), nrow = nrows)
+  y <- rowSums(x)
+  weights <- seq(from = 1, to = ncols)
+
+  test <- function(tm) {
+    names <- paste0("f", 1:ncols)
+    xy <- xgb.DMatrix(data = x, label = y, feature_weights = weights)
+    params <- list(colsample_bynode = 0.4, tree_method = tm, nthread = 1)
+    model <- xgb.train(params = params, data = xy, nrounds = 32)
+    importance <- xgb.importance(model = model, feature_names = names)
+    expect_equal(dim(importance), c(ncols, 4))
+    importance <- importance[order(importance$Feature)]
+    expect_lt(importance[1, Frequency], importance[9, Frequency])
+  }
+
+  for (tm in c("hist", "approx", "exact")) {
+    test(tm)
+  }
+})

doc/parameter.rst

@@ -115,10 +115,9 @@ Parameters for Tree Booster
     'colsample_bynode':0.5}`` with 64 features will leave 8 features to choose from at
     each split.
 
-    On Python interface, when using ``hist``, ``gpu_hist`` or ``exact`` tree method, one
+    Using the Python or the R package, one can set the ``feature_weights`` for DMatrix to
-    can set the ``feature_weights`` for DMatrix to define the probability of each feature
+    define the probability of each feature being selected when using column sampling.
-    being selected when using column sampling.  There's a similar parameter for ``fit``
+    There's a similar parameter for ``fit`` method in sklearn interface.
-    method in sklearn interface.
 
 * ``lambda`` [default=1, alias: ``reg_lambda``]