A DENCLUE based approach to neuro-fuzzy system modeling

In order to solve the problems of difficulty to determine the number of partitions and rule redundancy in neuro-fuzzy system modeling, this paper presents a new approach based on DENCLUE using a dynamic threshold and similar rules merging (DDTSRM). By introducing DDT, which uses a dynamic threshold rather than a global one in merging density-attractors in DENCLUE, our approach is good at determining the number of partitions because DDT does not depend on input parameters. Additionally, the modeling performance is improved for DDT can find arbitrary shape and arbitrary density clusters. After structure identification we merge similar rules by considering similarity measures between fuzzy sets. Finally, BP method is used to precisely adjust the parameters of the fuzzy model. For illustration, we applied DDTSRM to a nonlinear function and Box and Jenkins system. Experimental results show that DDTSRM is effective to solve the problems with a good performance.