On designing of flexible neuro-fuzzy systems for classification
Krzysztof Cpałka, Leszek Rutkowski · 2007
Abstract:- In the paper we propose a new method for designing flexible neuro-fuzzy systems applied to classification. The systems are characterized by weighted triangular norms describing the importance of linguistic variables and rules. An algorithm for complexity reduction of such systems is developed and simulation results are presented. Key-Words:- neuro-fuzzy systems, triangular norms, algorithm for reduction of neuro-fuzzy systems 1.