Bi-criteria Genetic Selection of Bagging Fuzzy Rule-based Multiclassification Systems

Previously we proposed a scheme to generate fuzzy rule-based multiclassification systems by means of bagging, mutual information-based feature selection, and a multicriteria genetic algorithm (GA) for static component classifier selection guided by the ensemble training error. In the current contribution we extend the latter component by the use of two bi-criteria fitness functions, combining the latter error measure with the selected ensemble likelihood. A study on four popular UCI datasets with different dimensionalities is conducted in order to analyze the accuracy-complexity trade-off obtained by the two GAs, the initial fuzzy ensemble and a single fuzzy classifier. Keywords-- Bagging, feature selection, fuzzy rule-based multiclassification systems, genetic selection of individual classifiers, multicriteria genetic algorithm.

• Type: Journal
• Year: 2009
• Where: EUSFLAT
• Authors: Krzysztof Trawinski, Arnaud Quirin, Oscar Cordón
• Keywords: EUSFLAT 2009, Fuzzy Logic, Fuzzy Rule-based Multiclassification, Multicriteria Genetic Algorithm, Rule-based Multiclassification Systems