B—P网络泛化性能的改善

在神经网络的训练过程中存在“过度吻合”的现象，即训练样本的误差已达到非常小的一个值，但是非训练样本的误差非常大，造成神经网络的泛化性能不好。本文说明了泛化性能与隐层节点数的关系，并提出了通过改变性能函数来改善B－P网络的泛化性能的方法。

Improvement of BP Network's Generalization

One of the problems that occurs during neural network training is called overfitting. The error on the training set is driven to a very small value, but when new data is presented to the network the error is large. It will result in poor generalization. The relationship between generalization and the sum of the hidden nodes is described. A method by modifying performance function to improve generalization of B P network is also suggested.