A self-tuning of fuzzy reasoning by modifying learning coefficients

Recently, various methods to tune fuzzy rules have been proposed. In such algorithms, the time for computing suitable parameters of MSFs to match rules to the supervised data is not so short. And the generalization capability to express the behavior of a plant can not be realized sufficiently by their tuned rules. To improve such difficulties, premises and consequences of the rules are tuned by using the delta rules known as a learning rule of the neural network where the membership functions are given as triangles. The tuning is implemented by adjusting the lengths of the bases of the triangles so as to minimize the reasoning errors. In the present paper, we propose to apply piecewise linear MSFs to a self-tuning of fuzzy reasoning. Comparing with those by the conventional triangular MSFs, its adaptability is expansible. Also, for constructing such inference rules, the learning time can be reduced. Besides, a method to modify learning coefficients is proposed, which will serve to avoid the suboptimal solution.

Recommendations

Cited in

(4)

This page was built for publication: A self-tuning of fuzzy reasoning by modifying learning coefficients

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4378785)