Systems, Man, and Cybernetics

 

Home

• [P1062] A Wavelet-Based Recurrent Fuzzy Neural Network Trained With Stochastic Optimization Algorithm
  Author(s): Ahmad Taha Abdulsadda, Kamran Iqbal
  
  

  Abstract
  this paper presents a wavelet-based recurrent fuzzy neural network (WRFNN) trained with stochastic search-based adaptation algorithm. A WRFNN model proposed by Cheng, Jian and Cheng in 2004 [1], was modified by application of Simultaneous Perturbation Stochastic Analysis (SPSA) method. Generally, all neural networks with its modified structure still suffer from the local minimum problem. The SPSA algorithm was shown to be a stable global optimization technique that is applicable to WRFNN models with demonstrated computational advantages over other optimization algorithms.

  
   

Full Text (PDF)