Interpretable Selection and Visualization of Features and Interactions Using Bayesian Forests

TL;DR

This paper introduces a Bayesian forest-based method that enhances interpretability in machine learning by selecting relevant features and interactions, providing visual insights while maintaining competitive predictive accuracy.

Contribution

The paper presents the Selective Bayesian Forest Classifier, a novel approach combining classification, feature selection, interaction detection, and visualization using Bayesian networks and MCMC sampling.

Findings

Performs well on classification benchmarks in various dimensions.

Effectively identifies relevant features and interactions.

Provides visual tools for interpretability.

Abstract

It is becoming increasingly important for machine learning methods to make predictions that are interpretable as well as accurate. In many practical applications, it is of interest which features and feature interactions are relevant to the prediction task. We present a novel method, Selective Bayesian Forest Classifier, that strikes a balance between predictive power and interpretability by simultaneously performing classification, feature selection, feature interaction detection and visualization. It builds parsimonious yet flexible models using tree-structured Bayesian networks, and samples an ensemble of such models using Markov chain Monte Carlo. We build in feature selection by dividing the trees into two groups according to their relevance to the outcome of interest. Our method performs competitively on classification and feature selection benchmarks in low and high dimensions, and includes a visualization tool that provides insight into relevant features and interactions.