Research On Obtaining Weight From Weighted Fuzzy Rules

We can learn a set of fuzzy production rules which are learned from training examples have poor reasoning accuracy with respect to the training examples. This paper proposes an approach to refine the fuzzy production rules, which assigns local weights to propositions of fuzzy production rules by using a linear program, We can improve reasoning accuracy by some technology of optimization. This paper discusses a group of weighted fuzzy production rules and their corresponding reasoning mechanism. The group of weighted fuzzy production rules is mapped to a feed-forward neural network in which connection weights are just the local and global weights in weighted fuzzy production rules. Then these weights can be acquired by training the neural network using the gradient-decent technique. From above articles, we present three improved training algorithms, they thought of accuracy and efficiency of fuzzy neural network, making it more efficient implement.