股市时间序列的非线性分析  

Nonlinear Analysis of Stock Market Time Series

作　　者：胡广生[1] 臧保将[1] 商朋见[1] 

机构地区：[1]北京交通大学理学院,北京100044

出　　处：《北京交通大学学报》2006年第6期60-64,共5页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：科技部"973"基金项目(2004CB318005)

摘　　要：利用自相关函数对恒生指数日收盘价时间序列中的自相似性进行了初步的实证和理论分析.在此基础上,运用Husrt指数,进行了R/S(Re_scale Range Analysis)分析,以侦测收盘价的长程相关性.最后,对恒生指数日收盘价时间序列进行相空间重构,对其关联维数以及Kolmogorov熵进行了计算.研究结果表明,恒生指数日收盘价变化具有混沌等非线性性质.这一结论对股票市场理论建模、短时预测和管控策略的制定具有重要的意义.We do the initial identification and the theoretical analysis with the autocorrelation function towards the time series of HSI daily closing quotation, On the basic of it, with the Hurst index, we do the R/S analysis for the time series HSI daily closing quotation to examine the long range dependence. At last,we do the reconstruction of the phase space and compute its correlation dimension and Kolmogorov entropy. The research results show the change of HSI daily closing quotation has the chaos and other nonlinear features. The conclusion has important meaning for the stock market theoretical modelling, short forecast and making the control strategic plan.

关 键 词：恒生指数 自相似性 自相关函数 重构相空间 R/S分析 关联维数 Kolmogorov熵 

分 类 号：O211.61[理学—概率论与数理统计]