From 68a40c790bebfb47f6d9ac16fc4304eaf714bbf3 Mon Sep 17 00:00:00 2001 From: James Lamb Date: Tue, 29 Dec 2020 00:25:39 +0000 Subject: [PATCH] [docs] add doc on min_data_in_leaf approximation (fixes #3634) (#3690) * [docs] add doc on min_data_in_leaf approximation (fixes #3634) * Fix capital letter Co-authored-by: Nikita Titov --- docs/Parameters.rst | 2 ++ include/LightGBM/config.h | 1 + 2 files changed, 3 insertions(+) diff --git a/docs/Parameters.rst b/docs/Parameters.rst index 004e87e96532..235dbb9040a2 100644 --- a/docs/Parameters.rst +++ b/docs/Parameters.rst @@ -284,6 +284,8 @@ Learning Control Parameters - minimal number of data in one leaf. Can be used to deal with over-fitting + - **Note**: this is an approximation based on the Hessian, so occasionally you may observe splits which produce leaf nodes that have less than this many observations + - ``min_sum_hessian_in_leaf`` :raw-html:`🔗︎`, default = ``1e-3``, type = double, aliases: ``min_sum_hessian_per_leaf``, ``min_sum_hessian``, ``min_hessian``, ``min_child_weight``, constraints: ``min_sum_hessian_in_leaf >= 0.0`` - minimal sum hessian in one leaf. Like ``min_data_in_leaf``, it can be used to deal with over-fitting diff --git a/include/LightGBM/config.h b/include/LightGBM/config.h index f1678c478394..b8dd3817407c 100644 --- a/include/LightGBM/config.h +++ b/include/LightGBM/config.h @@ -274,6 +274,7 @@ struct Config { // alias = min_data_per_leaf, min_data, min_child_samples // check = >=0 // desc = minimal number of data in one leaf. Can be used to deal with over-fitting + // desc = **Note**: this is an approximation based on the Hessian, so occasionally you may observe splits which produce leaf nodes that have less than this many observations int min_data_in_leaf = 20; // alias = min_sum_hessian_per_leaf, min_sum_hessian, min_hessian, min_child_weight