作　　者： ; ; ;

机构地区： 华南理工大学工商管理学院

出　　处： 《广东金融学院学报》 2011年第1期65-76,89,共13页

摘　　要： 通过引入滑动窗口技术对Kantelhardt(2002)提出的多重分形消除趋势波动分析法(MF-DFA)加以改进,对中信标普公司推出的6种纯风格资产指数日收盘价格波动序列的分形特征进行了研究。实证结果表明,滑动窗口技术能有效减少因分割连接点处的不连续性而产生的伪波动误差,具有更高的精度;风格资产价格波动原始序列具有持久性,相位重构序列具有反持久性,且相位重构序列的多重分形特征显著弱于原始序列的多重分形特征,说明风格资产价格波动序列的持久相关性是形成多重分形特征的主要原因,表现出相关多重分形特征;价值、成长型比规模型风格资产具有更强的多重分形特征。 This paper analyzed the fractal characteristics of stock style assets returns by introducing a sliding window technology to the multifractal detrended fluctuation analysis （MF-DFA） proposed by Kantelhardt （2002）. Six kinds of pure stock style assets index daily return volatility characteristics being introduced by Standard ＆ Poor Company were studied. The empirical results showed that： sliding window technology can effectively reduce for segmentation join -point place of discontinuity and generate pseudo fluctuation error with higher precision ; the original style asset series of the style asset index is persistent, phase reconstruction series have anti -persistence and multi-fractal characteristics of phase reconstruction series were significantly weaker than the multi -fractal characteristics of original series. It showed enduring relevance of style asset index return series was the main reasons of multi -fractal characteristics; value growth has stronger muhifractal characteristics than the size style assets.

关 键 词： 风格资产 价格波动 滑动窗口 广义 指数 多重分形强度

领　　域： [经济管理]