From adcd8ea7c6723e2297e35d52a56a2fa3dcc661bf Mon Sep 17 00:00:00 2001 From: Nan Zhu Date: Thu, 16 May 2019 21:17:57 -0700 Subject: [PATCH] Update xgboost4j_spark_tutorial.rst (#4476) --- doc/jvm/xgboost4j_spark_tutorial.rst | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/doc/jvm/xgboost4j_spark_tutorial.rst b/doc/jvm/xgboost4j_spark_tutorial.rst index d104e78c3..b25ab8324 100644 --- a/doc/jvm/xgboost4j_spark_tutorial.rst +++ b/doc/jvm/xgboost4j_spark_tutorial.rst @@ -192,9 +192,9 @@ Example of setting a missing value (e.g. -999) to the "missing" parameter in XGB .. note:: Using 0 to represent meaningful value -Due to the fact that Spark's VectorAssembler transformer only accepts 0 as a missing values, this one creates a problem when the user has 0 as meaningful value plus there are enough 0's to use SparseVector (However, In case the dataset is represented by a DenseVector, the 0 is kept) + Due to the fact that Spark's VectorAssembler transformer only accepts 0 as a missing values, this one creates a problem when the user has 0 as meaningful value plus there are enough 0's to use SparseVector (However, In case the dataset is represented by a DenseVector, the 0 is kept) -In this case, users are also supposed to transform 0 to some other values to avoid the issue. + In this case, users are also supposed to transform 0 to some other values to avoid the issue. Training ========