摘要
提出一种基于二进正交小波变换和AR-LSSVM方法的非平稳时间序列预测方案.首先利用Mallat算法对非平稳时间序列进行分解和重构,分离出非平稳时间序列中的低频信息和高频信息;然后对高频信息构建自回归模型,对低频信息则用最小二乘支持向量机进行拟合;最后将各模型的预测结果进行叠加,从而得到原始序列的预测值.研究结果表明,该方法不仅能充分拟合低频信息,而且可避免对高频信息的过拟合.
A non-stationary time series prediction method using the wavelet analysis and AR-LSSVM is proposed. By wavelet decomposition and reconstruction, the non-stationary time series are decomposed into a low frequency signal and several high frequency signals. The high frequency signals are predicted with auto-regression models, and the low frequency is predicted with least square support vector machines. The prediction result of the original time series is the superimposition of the respective prediction. This new method avoids the over-fitted for high frequency signals, and adequately fits the low signal of the non-stationary time series, so better predicting performance can be obtained. Experiments show the effectiveness of the predicting method.
出处
《控制与决策》
EI
CSCD
北大核心
2008年第3期357-360,共4页
Control and Decision
基金
甘肃省自然科学基金项目(ZS022-A25-039)
关键词
小波变换
非平稳时间序列
最小二乘支持向量机
自回归
预测
Wavelet transform
Non-statlonary time series
Least squares support vector machines
Auto-regression
Prediction
