Automatic fuzzy rule extraction based on particle swarm optimization

The extraction of fuzzy rules is always a difficult problem to fuzzy system. We have proposed a pruning algorithm to optimize fuzzy neural network based on particle swarm optimization algorithm. It can evolve both the fuzzy neural network´s topology and weighting parameters. In a real problem, it can automatically obtain the near-optimal structure of fuzzy neural network according to the requirements. The experiment has proved that the method is applicable and efficient.