CAJ | 학술논문

目前股票市场长记忆性检验和建模方法，不能很好地消除短期记忆的影响，针对这一问题，本文提出寻找序列的突变点，通过将序列分解为只包含长记忆性部分和不包含长记忆性部分的序列分解技术，来排除短期记忆的影响。对上证指数和深圳成分指数收益率波动的长记忆性进行实证研究发现，将序列分解以后进行长记忆性检验，不仅可以得出长记忆性检验更为精确的结论，同时可以检验序列分解过程的效果。
목전고표시장장기억성검험화건모방법，불능흔호지소제단기기억적영향，침대저일문제，본문제출심조서렬적돌변점，통과장서렬분해위지포함장기억성부분화불포함장기억성부분적서렬분해기술，래배제단기기억적영향。대상증지수화심수성분지수수익솔파동적장기억성진행실증연구발현，장서렬분해이후진행장기억성검험，불부가이득출장기억성검험경위정학적결론，동시가이검험서렬분해과정적효과。
The research methods for verifying and modeling the long memory on the stock market in recent years can not remove the influence of the short memory, how to remove the influence of the short memory becomes an important research. In this paper, we use ICSS algorithm to seek the breakpoints of stock market volatility series. The influence of the short memory is removed by decomposing the series into two parts： one only including long memory and the other not including long memory. We then use the rescale R/S to examine two series which have been decomposed. The result shows the success in decomposition process.