Interpretable decision trees through MaxSAT

Josep Alòs; Carlos Ansótegui; Eduard Torres

doi:10.1007/s10462-022-10377-0

Interpretable decision trees through MaxSAT

Artif Intell Rev. 2022 Dec 27:1-21. doi: 10.1007/s10462-022-10377-0. Online ahead of print.

Authors

Josep Alòs^#¹, Carlos Ansótegui^#¹, Eduard Torres^#¹

Affiliation

¹ Logic & Optimization Group (LOG), University of Lleida, Lleida, Spain.

^# Contributed equally.

Abstract

We present an approach to improve the accuracy-interpretability trade-off of Machine Learning (ML) Decision Trees (DTs). In particular, we apply Maximum Satisfiability technology to compute Minimum Pure DTs (MPDTs). We improve the runtime of previous approaches and, show that these MPDTs can outperform the accuracy of DTs generated with the ML framework sklearn.

Keywords: Decision trees; Explainable AI; MaxSAT.