A combined method for error and complexity reduction in fuzzy rule-based classification

Abstract

The question how to manage the contradictive requirements of accuracy and compactness in classification systems remains an important question in machine learning and data mining. This paper proposes a approach that belongs to the domain of fuzzy rule-based classification and uses the method of rule granulation for error reduction and the method of rule consolidation for complexity reduction. The cooperative nature of those methods - the rules are split in a way that makes efficient rule consolidation feasible, rule consolidation is capable of further error reduction - is demonstrated in a number of experiments with nine benchmark classification problems, confirming the robustness of the proposed approach.