A simplified learning algorithm for interval type-2 fuzzy neural network

This paper is devoted to the learning problem for the interval type-2 fuzzy neural network. The type-reduced set of the proposed neural network is firstly estimated by the linear combination of boundary type-1 fuzzy logic systems, and then the corresponding output estimation error is analyzed. Finally, a novel risk function is represented and a simplified back propagation learning algorithm is developed which can largely relieve the computation burden.