判别混沌与噪声的局部可变神经网络方法

Local Variable Weight Neural Network of Identifying Choas and Noise

为了解决多对单类型的分类问题（ 尤其是无限多对单） ，提出了一种处理多对单问题的新型神经网络模型———局部可变权值神经网络，并将其应用于混沌信号和噪声的判别．局部可变权值神经网络学习算法与传统神经网络类似，但在工作时可根据网络输入来调整神经网络局部权值．从最后分类结果看，无论是在学习时间还是在分类精度上。

The distinction between chaos and noise is made by classification where one to one refine to one characteristic to one mode. It can be solved by the neural network model. A refined model is proposed to resolve the many characteristics to one mode problem. The local variable weight neural network is used to distinguish chaos and noise. Improvements are made for the learning time and classification.