In this paper, a clustering-based algorithm is proposed for automatically constructing a multi-input fuzzy model where only the input-output data of the identified system are available. The proposed clustering method can automatically yield the number of clusters and its associated cluster centers from the input training data. While the features of training data are extracted by the proposed clustering method, the valuable information on the initial structure of the Sugeno-type fuzzy inference system is built up. For testing the performance of the proposed system modeling method, two wellknow examples from the literature and one real-world data set from the Taiwan's stock market are used to illustrate the validity of the proposed fuzzy system design procedure.
Relation:
International Journal of Advancements in Computing Technology 3(11), pp.394-401
