应用经验模态分解的上海股票市场价格趋势分解及周期性分析  

Trend-Cycle Decomposition and Cycle Analysis of Stock Price in Shanghai Stock Market Using Empirical Mode Decomposition

作　　者：秦宇[1] 

机构地区：[1]哈尔滨工程大学水声工程学院,黑龙江哈尔滨150001

出　　处：《中国管理科学》2008年第S1期219-225,共7页Chinese Journal of Management Science

摘　　要：文章以经验模态分解方法对上海股票市场价格日序列进行了趋势分解和周期性的波动分析。得到了趋势项和几个不同平均周期的非周期循环分量。实证显示,在趋势分解效能上,经验模态分解与Hodrick-Prescott滤波器性能相近。经验模态分解的结果支持了R/S分析对沪市长程相关性的考察结论。得到的各分量的波动对宏观的重大事件有不同程度的反应,这意味着这些分量的波动反映了股市的各种周期规律。平均周期为40天的分量对股市重要转折点有敏感的预期,其最长预测时间符合非线性动力学的理论,这个分量可以被分为波动特征不同的两个阶段。分析结果表明,2005年中国股权分置改革是沪市股票价格周期波动特征改变的过渡阶段。本文引入的经验模态分解方法提供了一种有用的股票时间序列分析工具,建立基于此结果的新的理论框架将是非常有意义的工作。In this paper,Empirical Mode Decomposition(EMD)was applied to analys the daily stock price series in Shanghai stock market.The trend component and non-periodic cycle components with different average cycle of the log-price series were found.The empirical analysis has showed that EMD had the similar effectiveness as Hodrick -Prescott filter in dealing with the problem of trend-cycle decomposition.The result represented the same long term persistence characteristic as that found by R/S Analysis.Several mode components decomposed by EMD had the different reflections to the macroeconomic events,which imply that these components can reveal different rules of fluctuations in Shanghai stock market.The component with an average cycle of 40 days can predict the turnover points sensitively and this maximum predictable time conforms to the nonlinear dynamics theory.We found that reform of the shareholder structure beginning in 2005 is a transition region on this component which can be divided into two areas according to evident different fluctuation characteristics in each area.EMD had worked as a useful tool for analysis of stock price time series in this paper.The research to develop a new theoretic framework based on the results revealed by EMD is a very valuable work.

关 键 词：经济周期 非周期循环 经验模态分解 波动特性 模态分量 

分 类 号：F832.51[经济管理—金融学] F224