小波包与混沌集成的股票市场预测新方法.pdfVIP

7 1 () Vo.l 7 No. 1 2 0 0 5 1 JO URNAL OF H IT ( SOCIAL SCIEN CE S ED ITION ) Jan. , 2005 : 2005) 01- 0064- 05 (天津大学, 天津 300072) : 基于小波包变换和混沌理论提出了一种股票市场建模及其预测 新方法, 既能刻划时间序列 规律,又能捕捉混沌状态 特征首先, 应用小波包变换对上证综指和深证成指日收益率序列进行三层分解, 分别得到第三层从低频到高频八 个频率成分 时序, 并在此基础上作进一步分析, 结果表明中国股市存在混沌 特性;然后, 应用混沌理论分别建立从低频到高频八 个时序 预测模型,分别对八 个时序进行预测; 最后, 基于 小波包理论对混沌模型预测 结果予以重构, 实现对原始收益率序列 预测与现有方法比较, 该方法具有较 高 精度和较大 应用范围 : 小波包变换; 股票; 混沌; 预测 : F830. 91 : A A new stock m arket predicting m ethod based on in tegration of wavelet packet and chaos theory YIN Guang- e,i ZHENG P i- e (T ianjinUniversity, T ianjin 300072, China) A b stract: A ne method is proposed to predict the stockmarketbased on aveletpacket transformation and chaos theory, hich can not only describe a tmi e series but also capture the features of the chaos. F irst, by using the avelet packet transformation, ShanghaiComposite Index ( SCI) and Shenzhen Sub- Index ( SSI) daily gain rate series are de composed respectively into eight tmi e seriesparts from lo frequency to high frequency, and further analysis of decom position indicates that there exists a chaos in Chinas stockmarket. Then, by using the chaos theory, the chaotic mod els are established to predict the eight tmi e series. Finally, the predicted results of chaotic models are reconstructed based on avelet packet theory. By doing so, the prediction of the original SCI and SSI daily gain rate series can be obtained. Smi ulation results demonstrate that the proposedmethod is of high precision and has ide potential applica tions. K ey w ords: aveletpacket transformation; stockmarket; chaos t