Random sampling fuzzy c-means clustering and recursive least square based fuzzy identification

In this paper, a new method of fuzzy identification based on fuzzy clustering and recursive least square is proposed. The membership degree of each given pattern is calculated by using fast fuzzy clustering algorithm and the consequent parameters are identified by recursive least square. It is shown that the computer CPU time has been greatly saved compared with fuzzy c-means clustering method. A numerical example is given at the end of the paper to demonstrate the applicability and validity of the proposed method