Tutorial on Feature Importance

In this tutorial, I discussed various feature importance methods for regression/classification setting as how to find important input features which drive the target feature. I also discussed the advantages, and disadvantages of these methods. A regression problem has been taken into consideration to determine the important factors which drive the regression target.

Methods in consideration:

1. Random Forest/Gradient Boost Regression Feature Importance
2. Permutation Feature Importance
3. Dependency Plots
4. SHAP: SHapley Additive exPlanations
   1. Summary Plot and Partial Dependency Plot have been discussed and were used to determine the important factors driving the target variable.