Multiobjective Formulations of Fuzzy Rule-Based Classification System Design
Hisao Ishibuchi, Yusuke Nojima · 2005
We examine several formulations of fuzzy rule selection for the design of fuzzy rule-based classification systems in our two-stage approach. The first stage is heuristic rule extraction where a large number of candidate rules are extracted. The second stage is evolutionary rule selection where fuzzy rule-based systems are constructed by choosing a small number of candidate rules. Rule selection is formulated as single-, two-, and three-objective optimization problems using an accuracy measure and two complexity measures.