Learning Qualitatively Diverse and Interpretable Rules for Classification

TL;DR

This paper introduces a method to identify multiple distinct and accurate models for a dataset, emphasizing the discovery of simpler, interpretable classifiers when data supports multiple solutions.

Contribution

The work presents a novel approach to find a maximal set of diverse, accurate models, improving interpretability and understanding of data.

Findings

When multiple accurate classifiers exist, simpler models are often recovered.

The method effectively identifies a set of diverse models.

It enhances interpretability by focusing on simpler classifiers.

Abstract

There has been growing interest in developing accurate models that can also be explained to humans. Unfortunately, if there exist multiple distinct but accurate models for some dataset, current machine learning methods are unlikely to find them: standard techniques will likely recover a complex model that combines them. In this work, we introduce a way to identify a maximal set of distinct but accurate models for a dataset. We demonstrate empirically that, in situations where the data supports multiple accurate classifiers, we tend to recover simpler, more interpretable classifiers rather than more complex ones.