股票市场收益的长期记忆特征对于系统非线性结构的确定以及市场有效性的研究具有重要的意义.针对上海和深圳的周和日收益序列,采用非线性估计方法提高R/S系列分析估计H参数的精确度.并用ARFIMA模型对沪深股市的收益率的长期记忆性进行了检验,根据分段检验的结果,得出一些中国证券市场有效性的结论.
Analysis for the characteristic of long-memory in stock market returns has important sense in research of market effectiveness and framework determination of nonlinear system. According to the index of weekly and daily returns of Shanghai and Shenzhen, the paper utilizes the non-linearity estimation method to enhance the precision of the R/S series analysis on estimating the H pro'refuter, and checks the longmemory in stock market returns of Shanghai and Shenzhen using ARFIMA model. According to the parti- tioned examination results, some valid conclusions of securities market of China have been drawn.