Recurrent Interval Type-2 Fuzzy Neural Network Using Asymmetric Membership Functions
Ching‐Hung Lee, Tzu-Wei Hu · InTech eBooks · 2008
This article has introduced a novel recurrent interval type-2 fuzzy neural network with asymmetric membership functions, which utilizes Lyapunov stability theorem to prove the stability of the system. The novel RiT2FNN-A use the interval asymmetric type-2 fuzzy sets implements the FLS in a five-layer neural network structure which contains four layer forward network and a feedback layer. According to the Lyapunov theorem and gradient descent method, the convergence of RiT2FNN-A is guaranteed and the corresponding learning algorithm is derived. Moreover, the RIT2FNN-A capability to temporarily store information allowed us to extend the application domain to include temporal problem. In application, We have found that the proposed RiT2FNN-A can use a smaller network structure and a small number of tuning parameters than the feed-forward fuzzy neural networks to obtain similar or better performance. It can successfully also approximate to a dynamic system mapping as accurately as desired.