Affiliations: Department of Computer Science and Intelligent Systems, Graduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan
Correspondence:
[*]
Corresponding author: Prof. Hisao Ishibuchi, Department of Computer Science and Intelligent Systems, Graduate School of Engineering, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan. Tel.: +81 72 254 9350; Fax: +81 72 254 9915; E-mail: [email protected]
Abstract: In this paper, we examine the application of evolutionary multiobjective optimization (EMO) algorithms to the design of fuzzy rule-based ensemble classifiers. An EMO algorithm is used to search for a large number of non-dominated fuzzy rule-based classifiers along the accuracy-complexity tradeoff surface. The accuracy of each fuzzy rule-based classifier is measured by the number of correctly classified training patterns while its complexity is measured by the number of fuzzy rules and the total number of antecedent conditions. An ensemble classifier is designed by combining non-dominated fuzzy rule-based classifiers. We examine the performance of ensemble classifiers through computational experiments on six benchmark data sets in the UCI machine learning repository. We also examine the diversity of individual fuzzy rule-based classifiers in each ensemble classifier.
