TSK interval type-2 fuzzy neural networks for chaotic time series prediction
Liang Zhao · 2010
This paper presents TSK interval type-2 fuzzy neural network (TSK IT2FNN)and its learning algorithm for chaotic time series prediction. First, The structure of TSK IT2FNN is decided using the hierarchical fuzzy clustering algorithm. Then its parameters of the precondition membership function and consequence weight are optimized using the gradient descent algorithm. Finally the effectiveness of IT2FNN and its learning algorithm are evaluated by using the Mackey-Glass chaotic time series. The simulation result shows that this proposed method is effective in the paper.