A Parallel Learning Algorithm for Fuzzy Neural Networks Based on Fuzzy Numbers Operations

摘要

Fuzzified neural network based on fuzzy number operations may be one of powerful fuzzified neural networks. We introduced a fuzzified neural network based on fuzzy number operations which approximates targets by choosing the shapes of the weights and the biases and calculating the left, mean and right values of each weight or bias. However, as we have pointed out, the learning algorithm is a little time-consuming. We propose a parallel learning algorithm for the fuzzified neural network in this paper. The parallel learning algorithm works well for the fuzzified neural network.