4 Model agnostic methods – local interpretability
This chapter covers:
- Characteristics of deep neural networks
- How to implement deep neural networks that are inherently black box
- Perturbation-based model agnostics methods that are local in scope such as LIME, SHAP and Anchors
- How to interpret deep neural networks using LIME, SHAP and Anchors
- Strengths and weaknesses of LIME, SHAP and Anchors
In the previous chapter, we looked at tree ensembles especially Random Forest models and learned how to interpret them using model agnostic methods that are global in scope such as Partial Dependence Plots (PDPs) and feature interaction plots. We saw that PDPs are a great way of understanding how individual feature values impact the final model prediction at a global scale. We were also able to see how features interact with each other using the feature interaction plots and also how they can be used to expose potential issues such as bias. They are easy and intuitive to understand but a major drawback of PDPs is that it assumes features are independent of each other. Higher order feature interactions can also not be visualized using feature interaction plots.