Improved Discrete Binary Harmony Search Algorithm for Fuzzy Classifier Design

One of the most important issues in the design of fuzzy classifiers is the fuzzy rule base formation. This paper presents an Improved Discrete Binary Harmony Search (IDBHS) algorithm based approach for fuzzy classifier design. The optimal parameters of the improved discrete binary harmony search based fuzzy classifier (IDBHS-fuzzy classifier) including fuzzy membership functions and structure of fuzzy rules are extracted from the training data by evolving both of them using DBHS simultaneously. Harmony search algorithm is conceptualized using the musical improvisation process of searching for a perfect state of harmony. In this paper the impact of constant parameter on discrete binary harmony search algorithm is discussed and a strategy for tuning this parameter is presented that enhances accuracy and convergence rate of IDBHS-fuzzy classifier. Two pattern recognition problems with different feature vector dimensions were used to demonstrate the effectiveness of the proposed classifier. The experimental results show that the performance of the DBHS-fuzzy classifier is comparable to or better than the Genetic Algorithm fuzzy classifier (GA-fuzzy classifier) as a conventional fuzzy classifier and the k-nearest neighbor (k-NN) classifier as a traditional classifier, while its convergence speed is considerably higher

Improved Discrete Binary Harmony Search Algorithm for Fuzzy Classifier Design Keywords:

Fuzzy classifier , Harmony search algorithm , Optimization of fuzzy parameters , Pattern classification