摘要
基于R/S分析的Hurst指数揭示了自然界系统可观测量的幂函数关系,并在证券市场波动的复杂性研究中得到应用,但存在计算效率问题·作者利用标准差时间序列改进Hurst指数计算方法,并应用于国内沪深股市研究·结果表明,改进后的Hurst指数在计算效率上得到较大提高,沪深股市收益率均不服从正态分布,Hurst指数大于0 5,在跨时间尺度的股价指数之间存在着相关性,市场具有分形结构特征·
Although Hurst index based on R/S rescaled range analysis has revealed the exponential relationship between observable quantities in the nature and the relationship has been applied to the research on the volatile complexity of stock market, there is still an efficiency problem in calculation. Taking advantage of the time series of standard deviation to improve the efficiency in calculation of Hurst index and making use of the improvement on the two domestic stock markets in Shanghai and Shenzhen, the results showed that the calculating efficiency of improved Hurst index is much higher than the original one. The earnings ratios on the two stock markets are both out of accord with normal distribution, with Hurst indices both shown higher than 05. It is found there is an autocorrelation between stock price indices in accordance to a scale of time spanned, and the markets are characterized by a fractal structure.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2003年第9期862-865,共4页
Journal of Northeastern University(Natural Science)
基金
中国博士后科学基金资助项目(2002031148).
关键词
HURST指数
标准差时间序列
证券市场
重标极差分析
分形结构
Hurst index
time series of standard deviation
securities market
rescaled range analysis
fractal structure
