摘要
在不同的经济发展时期,股票市场波动会呈现出不同的动力学特征.鉴于分形理论在描述股票价格波动特性时具有许多优势,应用多重分形消除趋势波动分析(MF-DFA)对日本七个经济时期以及中国股市自建立以来三个经济阶段的股票市场指数进行实证研究.结果显示:不同经济发展时期日中两国的股票市场均具有明显的多重分形特性;但各自不同的经济时期多重分形特性差异显著,且与当时经济发展的状况存在着一定联系.接着运用自组织特征映射(SOM)神经网络对日本七个经济时期股市的多重分形特性进行分类,验证了多重分形消除趋势波动分析(MF-DFA)可以较准确地刻画出不同经济时期股票市场的动力学特征.最后,通过对比日中两国不同时期股票市场的多重分形性,得出一些对中国经济发展有益的启示.
The stock market fluctuation appears different dynamical charactemstlcs m amerent ecouomy development periods. As fractal has lots of advantages when describing the property of the price fluctu- ations, based on the multifractal detrended fluctuation analysis （MF-DFA）, the empirical research were brought forward to Japan stock market indices of the seven economy periods and China stock market indices of the three economy periods respectively. The results show that all the indices of Japan and China stock market have obvious multifractal properties, which differ from each other significantly and have some relations with the different economy status. And then, self-organizing feature map （SOM） neural network categorized Japan stock market indices of the seven economy periods, which tested and verified that MF-DFA can describe the dynamic characteristics of stock market in different time accurately. At last, some beneficial implications for multifractal properties of Japan China economy development are and China stock markets. obtained by comparing the time-varying
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2013年第2期317-328,共12页
Systems Engineering-Theory & Practice
基金
国家留学基金委"国家建设高水平大学公派研究生项目"([2008]3019
[2009]3012)
关键词
经济时期
股票市场
波动
多重分形
神经网络
economy periods
stock markets
fluctuation
multifractal
neural network
