摘要
证券市场是个典型的非线性复杂系统。本文运用修正R/S分析法对我国基金风格资产收益单一分形的基本统计特征进行检验,并与经典R/S方法进行对比分析。研究结果表明:在日、周、月等三种时间标度下Hurst指数均显著大于0.5,表现为持久相关性特征,说明股市风格具有长记忆性;从经典R/S分析结果看,我国股市风格具有显著的分形结构特征,风格资产指数收益率序列具有长记忆性,不同风格资产的业绩具有不同的周期性。
Securities market is a typical complicated nonlinear system. This paper uses modified R/S analysis to inspect single fractal basis statistical characteristics in fund style asset returns, and compares the result with the classic R / S method. The results show that, in the day, week and month time scale Hurst exponent is significantly greater than 0.5, for the enduring relevance of the performance characteristics, indicating that the stock market style has a long memory characteristics. From the classic R/S analysis results, China's stock market style has significant fractal structure, style asset index return series has a long memory feature, and the performance of different assets has different styles of periodicity.
出处
《南方金融》
北大核心
2011年第8期61-64,37,共5页
South China Finance