How Interpretable and Trustworthy are GAMs?

TL;DR

This paper evaluates various GAM algorithms, revealing that tree-based GAMs offer the best balance of interpretability, accuracy, and fairness, making them the most trustworthy choice among GAMs.

Contribution

It provides a comprehensive comparison of GAM algorithms, highlighting the importance of inductive bias and identifying tree-based GAMs as the most reliable for interpretability and fairness.

Findings

Tree-based GAMs balance sparsity, fidelity, and accuracy effectively.

High feature sparsity in GAMs can lead to missing data patterns and unfairness.

Inductive bias significantly influences what interpretable models learn.

Abstract

Generalized additive models (GAMs) have become a leading modelclass for interpretable machine learning. However, there are many algorithms for training GAMs, and these can learn different or even contradictory models, while being equally accurate. Which GAM should we trust? In this paper, we quantitatively and qualitatively investigate a variety of GAM algorithms on real and simulated datasets. We find that GAMs with high feature sparsity (only using afew variables to make predictions) can miss patterns in the data and be unfair to rare subpopulations. Our results suggest that inductive bias plays a crucial role in what interpretable models learn and that tree-based GAMs represent the best balance of sparsity, fidelity and accuracy and thus appear to be the most trustworthy GAM.