PRINCIPAL COMPONENTS, B-SPLINES, AND FUZZY SYSTEM REDUCTION
John Yen, Reza Langari, Liang Wang · International Journal of Uncertainty Fuzziness and Knowledge-Based Systems · 1996
This paper proposes to use the method of principal components to reduce the dimensionality of input space and the B-splines to represent the membership functions of input variables. A model reduction strategy, which is based on Johansen’s optimality theorem, also suggested. The utility of this approach is illustrated using a fuzzy system modeling example.