Trying to make sense of black box machine learning models such as Ensemble models
This repository consists the data, code and a presentation highlighting how we can use different ways to understand and interpret Ensemble Models. The resources will cover:
- Concept behind different Interpretable Machine Learning methods
- Pre-processing the data for ensemble and deep learning models
- Understanding variable importance by permutation based method
- Partial dependency method
- Independent conditional expectation (ICE)
- Surrogate model
- LIME
- Game theoretic method - Shapley
- DALEX method