Tech Reports

ULCS-03-021

MERBIS: A Self-Adaptive Multi-Objective Evolutionary Rule Base Induction System

Christian Setzkorn and Ray C. Paton

Abstract

Supervised classification is a particular data-mining task that forms part of the knowledge discovery process. Its objectives are to extract accurate, comprehensible and interesting knowledge form data. However, many existing supervised classification approaches only focus on one of these objectives. This paper introduces a Multi-Objective Evolutionary Rule Base Induction System called MERBIS that is capable of producing trade-off solutions with regard to the accuracy and comprehensibility objectives. We utilise a superior accuracy measure, problem-tailored genetic operators and a self-adaptive mechanism that reduces the number of parameters. We compare MERBIS with several existing approaches for supervised classification on a number of benchmark data sets and show that it performs comparable while producing more comprehensible classifiers.

For each technical report listed here, copyright and all intellectual property rights remain with the respective authors. Copyright is effective from the year of publication in each case. By downloading a file from this page, you agree to use it only for purposes of research and scholarship. Any other use of this material or storage of it in any medium or its sale or distribution in any form is expressly forbidden without prior written permission from the authors concerned.