A Framework for Evolving Fuzzy Classifier Systems Using Genetic Programming

Afuzzy classifier systemframeworkis proposedwhich employsa tree-based representation for fuzzy rule (classifier) antecedents and genetic programmingfor fuzzy rule discovery. Sucha rule representation is employedbecauseof the expressivepowerandgenerality it endowsto individual rules. Theframeworkproposes accuracy-basedfitness for individualfuzzyclassifiers and employs evolutionary competition between simultaneouslymatchedclassifiers. Theevolutionary algorithm(GP)is therefore searchingfor compactfuzzy rule baseswhichare simultaneouslygeneral,accurateand co-adapted. Additional extensions to the proposed frameworkare suggested.

• Type: Conference
• Year: 2001
• Where: FLAIRS
• Authors: Brian Carse, Anthony G. Pipe
• Keywords: Artificial Intelligence, FLAIRS 2001, Fuzzy Rule, Programmingfor Fuzzy Rule, Sucha Rule Representation