基于熵算法的股票指数高频数据复杂度测算与评价  

The Study of CSI 300 Index's Complexity and Comparison of Model Efficiency Based on Entropy Algorithm

作　　者：温博慧[1] 袁铭[2] 侯笠 

机构地区：[1]天津财经大学经济学院,天津300222 [2]天津财经大学理工学院,天津300222

出　　处：《经济数学》2015年第1期19-25,共7页Journal of Quantitative Economics

基　　金：国家自然科学基金青年项目(71103126)

摘　　要：在日内高频环境下检验基于兼容法的柯尔莫哥洛夫熵、样本熵和模糊熵等复杂度测算方法对我国沪深300股票指数的测算效率,并运用筛选后的有效算法分阶段研究和比较了序列复杂度的变化过程与变化幅度.结果表明,模糊熵算法是一种更适用于我国沪深300股票指数的有效复杂度测算方法,其对相似容忍度的敏感性更低,测度值连续性更好.随时间推移,我国沪深300股票指数复杂度整体呈上升趋势,而相较于发达市场甚至周边新兴市场其复杂度偏低.This paper studied the high frequency data of the CSI 300 index,and examined the efficiency of complexity measures such as Kolmogorov entropy,sample entropy and fuzzy entropy in high frequency environment.By using the effective measurement,it compared the changing process and range of the complexity both before and after the subprime crisis.The results show that,compared with the Kolmogorov entropy based on the compatible method and sample entropy,fuzzy entropy is more suitable for measuring the CSI300index＇s complexity,which has the lower sensitivity to the similar tolerance and the better continuity of measure value.The CSI 300index＇s complexity is rising during the sample interval.However,the complexity during the crisis is far more less than the two other stages,and the complexity after the crisis is higher than that before the crisis.Compared with the developed markets and even some emerging markets,the CSI 300index＇s complexity is much lower.

关 键 词：沪深300股票指数 复杂度 kolmogorov熵 样本熵 模糊熵 

分 类 号：F830[经济管理—金融学]