Lab 6: Interpreting Black-Box Models (LIME, SHAP) and PCA

Lab 6: Interpreting Black-Box Models (LIME, SHAP) and PCA

Slides

The slides from this week can be found here.

Lab Overview

• Focus: Interpreting Black Box Models with LIME and SHAP
• Bonus: Principal Component Analysis (PCA)
• Date: October 6, 2025

Topics Covered

• Interpreting black box models
  • Local Interpretable Model-Agnostic Explanations (LIME)
  • Shapley values (SHAP), specifically waterfall, beeswarm, dependence, and force plots
  • Principal Component Analysis (PCA)

Code

This Jupyter notebook collates the code from today’s lab.