add note on missing value types for R

microsoft · Jan 15, 2021 · 302a71c · 302a71c
1 parent 3c3615d
commit 302a71c
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/docs/Parameters.rst b/docs/Parameters.rst
@@ -127,7 +127,7 @@ Core Parameters
 
       -  categorical features are used for splits as normal but are not used in the linear models
 
-      -  missing values must be encoded as ``np.nan`` (Python) or ``NA`` (CLI), not ``0``
+      -  missing values should not be encoded as ``0``. Use ``np.nan`` for Python, ``NA`` for the CLI, and ``NA``, ``NA_real_``, or ``NA_integer_`` for R
 
       -  it is recommended to rescale data before training so that features have similar mean and standard deviation
 

diff --git a/include/LightGBM/config.h b/include/LightGBM/config.h
@@ -152,7 +152,7 @@ struct Config {
   // descl2 = tree splits are chosen in the usual way, but the model at each leaf is linear instead of constant
   // descl2 = the linear model at each leaf includes all the numerical features in that leaf's branch
   // descl2 = categorical features are used for splits as normal but are not used in the linear models
-  // descl2 = missing values must be encoded as ``np.nan`` (Python) or ``NA`` (CLI), not ``0``
+  // descl2 = missing values should not be encoded as ``0``. Use ``np.nan`` for Python, ``NA`` for the CLI, and ``NA``, ``NA_real_``, or ``NA_integer_`` for R
   // descl2 = it is recommended to rescale data before training so that features have similar mean and standard deviation
   // descl2 = **Note**: only works with CPU and ``serial`` tree learner
   // descl2 = **Note**: ``regression_l1`` objective is not supported with linear tree boosting