Falling Rule Lists

TL;DR

This paper introduces a Bayesian framework for learning falling rule lists, a type of ordered classification model inspired by healthcare risk stratification, offering an alternative to greedy decision tree methods.

Contribution

It presents a novel Bayesian approach to learn falling rule lists, enabling more principled and potentially more accurate risk stratification models.

Findings

Effective in healthcare risk stratification scenarios

Provides a non-greedy alternative to decision tree learning

Demonstrates improved interpretability and accuracy

Abstract

Falling rule lists are classification models consisting of an ordered list of if-then rules, where (i) the order of rules determines which example should be classified by each rule, and (ii) the estimated probability of success decreases monotonically down the list. These kinds of rule lists are inspired by healthcare applications where patients would be stratified into risk sets and the highest at-risk patients should be considered first. We provide a Bayesian framework for learning falling rule lists that does not rely on traditional greedy decision tree learning methods.