added feature importance answer for boosting (#166)

* added feature importance answer for boosting * Update theory.md reverted --------- Co-authored-by: Roman Polyansky <[email protected]> Co-authored-by: Alexey Grigorev <[email protected]>
alexeygrigorev · May 6, 2024 · 15737bc · 15737bc
1 parent 111cc12
commit 15737bc
Showing 1 changed file with 5 additions and 1 deletion.
diff --git a/theory.md b/theory.md
@@ -719,7 +719,11 @@ Yes, different frameworks provide different options to make training faster, usi
 
 **Feature importance in gradient boosting trees  —  what are possible options? ‍⭐️**
 
-Answer here
+With CatBoost you can use implemented method get_feature_importance for getting SHAP values. https://arxiv.org/abs/1905.04610v1
+
+It allows to understand how excluding features helps to provide better results. Higher value is better.
+
+Also you can add random noise column to your data (with normal distribution), calculate feature importance and exclude all features below noise importance.
 
 <br/>