时间序列预测的动态神经网络方法

Prediction Method of Time Series Based on Dynamic Neural Network

时间序列的预测在经济和工程领域具有十分重要的意义。文中利用动态神经网络的特性,提出对时间序列进行预测的动态神经网络方法,并利用设计的动态神经网络对杜芬(Duffing)方程的响应时间序列进行预测,结果表明动态神经网络可以较好地对动态系统的响应时间序列进行预测。

The prediction of time series is very important in the economic and engineering fields. In this paper, makes use of the characteristic of dynamic neural network, and offer a method using dynamic neural network to predict time series.In the late,use this network to predict the response of Duffing equation. The consequence shows that the dynamic network may have a good predictive effect for the response of some dynamic system.