Some Interesting Blogs/Resource/Projects

2017-08-30 12:05:54 +0800

Feature Engineering

Encoding:

1. Visiting: Categorical Features and Encoding in Decision Trees. It compares 4 different encoding (Categorical, numeric, One-Hot, Binary) in Decision Tree. Plots are very interesting. Conclusion: 1. Don’t use One-Hot Encoding. 2. Categorical features with large cardinalities (over 1000): Binary. 3. Categorical features with small cardinalities (less than 1000): Numeric.