摘要
本文提出一种基于主成分分析(PCA)和动态神经网络的多变量时间序列预报方法,并对具体实例建立多变量时间序列模型。仿真实验结果表明该网络具有很强的学习能力和泛化能力,适合进行非线性时间序列预报。
This paper proposes a multivariable time series predicting method based on principal component analysis （PCA） and dynamic neural network. And also a model is established as a concrete example. The simulating result shows that the network has strong study feature and generalization, and can be adapted to predict nonlinear time series.
出处
《中国西部科技》
2009年第10期27-28,共2页
Science and Technology of West China
