Design of Neuro-Fuzzy Networks Based on Respective Input Space for Pattern Recognition

The design of neuro-fuzzy networks based on fuzzy respective input space for pattern recognition is introduced in this paper. The premise part of the rules of the proposed networks is realized by partitioning of the fuzzy respective input space. The respectively partitioned spaces express the rules of the networks. The consequence part of the rules is represented by polynomial functions. The coefficients of consequence part of the rules are learned by the back-propagation algorithm. And the proposed networks are optimized using real-coded genetic algorithms. A numerical example for pattern recognition is given to evaluate the validity of the proposed networks.