Fuzzy Similarity Measures for Signal Pattern Classification

Pattern classification is an important task for many practical systems. Many classifier systems rely on similarity measures to classify unknown patterns. Signal patterns are an interesting class of patterns exhibited in many sensorbased systems. In this paper we present three fuzzy similarity measures that can be used for signal pattern classification. We use the three fuzzy similarity measures in a signal pattern classification architecture that combines different intelligent techniques. Results of testing the performance of the signal pattern classifier using each of the proposed similarity measures demonstrate their important contribution to the overall classifier performance.