mation approach to fuzzy modeling was considered. The present paper formulates the task of developing a formal approach, which would enable analyzing fuzzy systems in terms of their capability to describe uncertainties of input information using interval membership functions. The discussed approach would allow to introduce the information factor for evaluating the quality of fuzzy models functioning using interval membership functions, and to increase the adequacy of the application area representation by a developed fuzzy model. The proposed information factor is a target function based on type-2 interval membership functions. The introduced target function optimizes the quantity of mutual information that is reflected from the inputs of a fuzzy model to its outputs. A technique for generating fuzzy type-2 models, which are optimal according to the given quality factor, and an algorithm for building an interval fuzzy model from experimental data and implementing the transition from regular to interval membership functions were introduced. An example of the calculations using this technique for computing the entropy estimation on a fuzzy model’s output is given.
Alan : Fen Bilimleri ve Matematik; Mühendislik
Dergi Türü : Uluslararası
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|