论文部分内容阅读
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，局部可变权值神经网络比传统神经网络都要好
In order to solve many pairs of single-type classification problems (especially infinite pairs of single), a new neural network model dealing with many-to-one problems is proposed, which is a local variable weight neural network, which is applied to chaotic signals And noise discrimination. Local variable weight neural network learning algorithm similar to the traditional neural network, but at work according to the network input to adjust the local weights of neural networks. From the final classification results, both in the study time and classification accuracy, the local variable weight neural network is better than the traditional neural network