基于R/S分析的股市风格分形特征研究

Research on Stock Market Style Fractals Based on Multifractal R/S Analysis

运用多重分形R/S分析法研究我国股票市场的风格分形结构特征,通过计算中信标普公司推出的6种纯风格资产指数在不同时间标度下的收益率序列Hurst指数与平均循环周期,发现我国股票市场风格存在统计自相似性、标度不变性、长记忆性,不同风格资产指数收益序列具有不同的平均循环周期等分形特征,这为基金公司、基金经理及时地把握股市风格的轮换规律,构建适度风格漂移策略以获取短期超额收益提供了决策参考与理论支持。

This paper analyzes the fractal structure characteristics of the Chinese stock market style based on muhifractal R/S analysis. By calculating the Hurst index and the average cycle of six kinds of style asset indices at different time scales, the paper finds that the existence of statistical self - similarity, scale invariance, long memory about the stock market styles, different style asset indices have different cycle fractal characteristics, which provides the investment decision - making and theoretical support in constructing moderate style drift strategy to get short - term excess returns for fund company and fund managers.