Skip to content

Latest commit

 

History

History
6 lines (6 loc) · 634 Bytes

File metadata and controls

6 lines (6 loc) · 634 Bytes

Categorical_Encoding_Experimentation

  1. This repo contains code for experimenting with categorical encoding - WoE, Catboost, Target encoder and many more.
  2. Many of these encoders such as Catboost, Target, Leave-one-out are very popular when it comes to encoding a categorical variable with very high cardinlaity.
  3. WoE is very popular for low/high cardinality for Binary classification problem especially with Logistic Regression.
  4. Binary and Ordinal encoder could be useful when encoding an ordinal variable.
  5. The train and test data used in these notebooks can be found here: https://www.kaggle.com/c/cat-in-the-dat/data/.