Fuzzy Radial Basis Function Neural Networks with Information Granulation and Its Genetic Optimization

This paper concerns Fuzzy Radial Basis Function Neural Networks with Information Granulation (IG- FRBFNN) and its optimization by means of the Hierarchical Fair Competition-based Parallel Genetic Algorithm (HFC-PGA). In the proposed network, the membership function of the premise part of fuzzy rules is determined by means of Fuzzy C-Means clustering. Also, we consider high-order polynomial as the consequent part of fuzzy rules which represent the input-output characteristic of subspace and the weighted Least Squares (WLS) learning is used to estimate the coefficients of polynomial. Since the performance of IG-RBFNN model is affected by some parameters such as a specific subset of input variables, the fuzzification coefficient of FCM, the number of rules and the polynomial order of the consequent part of fuzzy rules, we need the structural as well as parametric optimization of the network. In this study, the HFC-PGA is exploited to carry out the structural as well as parametric optimization of IG-based FRBFNN. The proposed model is demonstrated with the use of the chaotic Mackey-Glass time series data.