一类模糊神经网络

本文介绍了一个Ｆｕｚｚｙ神经网的三种基本形式，情况（１）是输入是实数，但权重和（或）偏差项是模糊的；情况（２）是输入是模糊的，但权重和偏差项是实数，情况（３）是输入模糊的，同时权重和（或）偏差项也是模糊的，本文集中研究情况（１），并主要讨论了情况（１）的模型性能（能力）而且得到一个较好的结果。

A TYPE OF FUZZY NEURAL NETWORK

This paper will introduce basically three types of a fuzzy neural network (FNN):(1)FNN 1 has real number input but the weight and/or bias terms can be fuzzy;(2)FNN 2 has fuzzy input but real number weights and bias terms;(3)FNN 3 has both fuzzy input and fuzzy weights and/or fuzzy bias terms.In this paper we concentrate on the case where the input signals are real numbers but the weights and bias terms are fuzzy we mainly discuss the modeling capabiliteies of FNN 1 and obtain a good resulf.