RUMC: A Rule-based Classifier Inspired by Evolutionary Methods

TL;DR

The paper presents RUMC, a novel rule-based classifier inspired by evolutionary methods, which improves classification accuracy over existing methods through innovative rule mutation techniques, validated on multiple datasets.

Contribution

Introduction of RUMC, a rule-based classifier that uses evolutionary-inspired rule mutation to enhance accuracy over previous rule aggregation methods.

Findings

RUMC outperformed 20 classifiers on 40 datasets.

RUMC effectively uncovers insights from complex data.

Demonstrated consistent improvement in classification accuracy.

Abstract

As the field of data analysis grows rapidly due to the large amounts of data being generated, effective data classification has become increasingly important. This paper introduces the RUle Mutation Classifier (RUMC), which represents a significant improvement over the Rule Aggregation ClassifiER (RACER). RUMC uses innovative rule mutation techniques based on evolutionary methods to improve classification accuracy. In tests with forty datasets from OpenML and the UCI Machine Learning Repository, RUMC consistently outperformed twenty other well-known classifiers, demonstrating its ability to uncover valuable insights from complex data.