摘要
本文以上证指数周收益率为研究对象,分别采用重标极差分析法和修正重标极差分析法,通过计算V统计量的值对其进行长期记忆性的检验。由于不能排除V统计量的值存在超出上侧分位点的可能性,本文进行了双侧检验,并分析了R/S分析法产生偏差的原因。得出上证指数周收益率时间序列并未表现出显著的长期记忆性的结论。
In this paper, we utilize rescaled range analysis and modified rescaled range analysis to test the return series of Shanghai stock exchange for Long-term memory. The null hypothesis of short-term memory is tested using V statistic. Our study when using modified rescaled range analysis is based on a two-tail test, because the exiseence of the values of modified rescaled range statistic which are beyond the bound of acceptance region that may lead to rejection of null hypothesis cannot be excluded absolutely. Moreover, we calculate the bias led by rescaled range analysis and analyze its cause of formation. According to our study, there is little evidence of long-term memory in return series of Shanghai stock exchange.
出处
《数理统计与管理》
CSSCI
北大核心
2006年第5期610-615,共6页
Journal of Applied Statistics and Management
