SHAP with CatBoostClassifier for Categorical Features | Python Tutorial

preview_player
Показать описание
Combining CatBoost and SHAP can provide powerful insight into your machine learning models. Especially, when you are working with categorical features.

With other modelling packages, we need to first transform categorical features using one-hot encodings. The problem is that each binary variable will have its own SHAP value. This makes it difficult to see the overall contribution of the original categorical feature.

Using Python, we will apply SHAP to a CatBoostClassifier. We will see the model's advantage over other libraries is that it can handle non-numerical data without transforming them. This means that SHAP values of a CatBoost model are easy to interpret as there will only be one SHAP value for each categorical feature.

🚀 SHAP Course 🚀
The first 20 people to use the coupon code "CATSHAP24" will get 100% off!

🚀 Free XAI Courses 🚀

🚀 GitHub Link with Code 🚀

🚀 Companion article with link to code (no-paywall link): 🚀

🚀 Useful playlists 🚀

🚀 Get in touch 🚀

🚀 Chapters 🚀
00:00 Introduction
01:40 Python Code
07:03 When to use CatBoost?
Рекомендации по теме
Комментарии
Автор

🚀 SHAP Course 🚀
The first 20 people to use the coupon code "CATSHAP24" will get 100% off!

adataodyssey
Автор

Hey! Thank you for the video.

Just a note: XGBoost now automatically deals with categorical features like Catboost. You just need to pass enable_categorical=True when creating the XGBClassifier!

FP-mgqk
Автор

Thank you! You clearly have a talent for explaining complex topics. At my university, it feels like the XAI course is covering almost exclusively topics from your channel! I appreciate your work and hope you find the time and enjoy making these videos! Best, Tim

nrthern_lights
Автор

How can we get shap value for each category in Catboost model like we get after one hot encoding in XGboost?

SnehaChoudhary-ng
welcome to shbcf.ru