NEW CRITERIA FOR RULE SELECTION IN FUZZY LEARNING CLASSIFIER SYSTEMS

Abstract

Designing an effective criterion for selecting the best rule is a major problem in the<br />process of implementing Fuzzy Learning Classifier (FLC) systems. Conventionally confidence<br />and support or combined measures of these are used as criteria for fuzzy rule evaluation. In this<br />paper new entities namely precision and recall from the field of Information Retrieval (IR)<br />systems is adapted as alternative criteria for fuzzy rule evaluation. Several different<br />combinations of precision and recall are redesigned to produce a metric measure. These newly<br />introduced criteria are utilized as a rule selection mechanism in the method of Iterative Rule<br />Learning (IRL) of FLC. In several experiments, three standard datasets are used to compare and<br />contrast the novel IR based criteria with other previously developed measures. Experimental<br />results illustrate the effectiveness of the proposed techniques in terms of classification<br />performance and computational efficiency.