Rule-Mining based classification: a benchmark study

TL;DR

This paper introduces an exhaustive rule-mining algorithm that creates interpretable, accurate local models by transforming data into rule-based features, matching the performance of traditional global classifiers.

Contribution

It presents a novel rule-mining approach that generates interpretable local features, enabling accurate local models comparable to standard global classifiers.

Findings

The method produces models as accurate as LR, SVM, RF, and GBT.

Generated rule-based features are human-understandable.

The approach offers a balance of interpretability and predictive power.

Abstract

This study proposed an exhaustive stable/reproducible rule-mining algorithm combined to a classifier to generate both accurate and interpretable models. Our method first extracts rules (i.e., a conjunction of conditions about the values of a small number of input features) with our exhaustive rule-mining algorithm, then constructs a new feature space based on the most relevant rules called "local features" and finally, builds a local predictive model by training a standard classifier on the new local feature space. This local feature space is easy interpretable by providing a human-understandable explanation under the explicit form of rules. Furthermore, our local predictive approach is as powerful as global classical ones like logistic regression (LR), support vector machine (SVM) and rules based methods like random forest (RF) and gradient boosted tree (GBT).