AN ALTERNATIVE APPROACH FOR GENERATION OF MEMBERSHIP FUNCTIONS AND FUZZY RULES BASED ON RADIAL AND CUBIC BASIS FUNCTION NETWORKS

The theoretically attractive fact that the radial basis function netwo rks can be interpreted as fuzzy systems is of small importance for pra ctical applications such as diagnosis and quality control with large n umbers of inputs or hidden neurons, due to the lack of transparency of the resulting fuzzy systems. A novel method for the generation of fuz zy classification systems based on radial basis function networks with restricted Coulomb energy learning is presented. The neural network a nd the learning algorithm are modified for easy hardware implementatio n by introducing cubic basis functions. The proposed methods are reste d with three application examples. The simulation results show the gen eration of compact, transparent fuzzy classification systems with good performance.