首先对神经网络应用于时间序列预测的方法进行了详细的介绍。在此基础上与传统预测方法进行了比较,接着概括分析了几种不同的神经网络结构用于预测的效果。指出由于神经网络独特的信息处理能力,使得它为一类高度非线性动态关系的时间序列预测提供了一条有效途径。最后,对神经网络在时间序序列预测领域的进一步研究提出了几点建议。 Firstly, the method of neural network applied to time series prediction is introduced in detail. Based on this comparison with the traditional prediction methods, the effects of several different neural network structures for prediction are summarized. It is pointed out that due to the unique information processing capability of neural network, it provides an effective way to predict the time series of a highly nonlinear dynamic relation. Finally, some suggestions are put forward for the further study of neural network in the field of time series prediction.