基于三种方法的股市重分形特性研究  

Multifractal analysis of stock market based on three methods

作　　者：蒋琼琼 杨毅恒[1] 黄静静[1] 

机构地区：[1]北京信息科技大学理学院

出　　处：《北京信息科技大学学报（自然科学版）》2016年第4期45-49,61,共6页Journal of Beijing Information Science and Technology University

基　　金：国家自然科学基金面上项目(61473325);北京市教委面上项目(KM201511232009)

摘　　要：对于股票市场的日收益率问题,采用时间序列的3种重分形分析法,即:重分形去趋势波动分析法、基于统计矩函数的重分形分析法和基于高高相关函数的重分形分析法分别进行研究,并应用上证和深证股票市场的日收益率序列对3种方法进行了数值实验。实验结果表明:上证和深证的股票市场时间序列均具有重分形特性。这一结果对于复杂的股票市场的行为研究具有重要意义,同时,股票市场的重分形特性在市场预测或风险管理中具有潜在的应用价值。Three kinds of multifractal analysis of time series is applied to solve the problem of stock daily return rate,namely multifractal detrended fluctuation analysis,multifractal analysis based on partition function, and multifractal analysis based on high-high correlation function. Numerical experiments are carried out on the three methods applying Shanghai and Shenzhen stock daily return rate. Experimental results show that the time series of Shanghai and Shenzhen stock market have multifractal characteristics. The results are of significance for the study of behaviors in the complex stock market. At the same time,the multifractal characteristics of the stock market have potential application value in the market forecast and risk management.

关 键 词：重分形分析 时间序列 股票市场 

分 类 号：O212[理学—概率论与数理统计]