Reference: POL 51 (W:2024)

Reference: POL 51 (W:2024) Sections 05-06

Extra examples and code for the interpretable machine learning methods introduced in the course with a variety of package implementations (i.e., IML, pdp, fastshap, vip, etc.). Also, some bonus (experimental) material using LOCO.

A brief script that demonstrates the superior predictive validity of local linear forests on Friedman's classic DGP.