Research on Fuzzy Neural Network Modeling Method Based on PSO Clustering Algorithm

Aim at the high dimensionality and rules extracting difficult issues of complex systems modeling, the theory of clustering analysis idea of pattern recognition system be introduced, proposes a method to extract fuzzy rules of the sample data based on POS Clustering algorithm.By analyzing sample data, the paper classify its clustering number and adjust class center, combined with the characteristics of the method, we establish a fuzzy neural network structure based on POS clustering algorithm. The fuzzy RBF network training algorithm used to adjust the membership function parameters and connection weights to complete the network parameter identification. The simulation shows that the method is suitable for the modeling of complex systems, and it has advantages of the high identification accuracy, fast convergence and rules self-extracting.