A new approach to TS fuzzy modeling using dual kernel-based learning machines

This paper proposes a novel approach for structure identification of TS fuzzy model using dual kernel-based learning machines. Firstly, a convenient kernel fuzzy C-means clustering algorithm is developed to partition the data set into several clusters. Secondly, a new kernel function which is free of parameter selection is utilized to locate support vectors in each cluster. Finally, the model structure is further simplified by a combination strategy for support vectors. The experimental results show that the resulting model has concise structure and good generalization ability, especially its performance is insensitive to initial clustering number.