主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2006, Vol. ›› Issue (3): 1-6.

• 论文 •    下一篇

基于小波分析的股市高频互相关研究

侯守国, 张世英   

  1. 天津大学管理学院, 天津, 300072
  • 收稿日期:2005-06-30 修回日期:2006-04-04 出版日期:2006-06-28 发布日期:2012-03-07
  • 基金资助:
    国家自然科学基金资助项目(70471050)

Research of Cross-Correlation of High-Frequency Data in Stock Markets with Wavelet Analysis

HOU Shou-guo, ZHANG Shi-ying   

  1. School of Management, Tianjin University, Tianjin 300072, China
  • Received:2005-06-30 Revised:2006-04-04 Online:2006-06-28 Published:2012-03-07

摘要: 以高频数据为对象,研究沪深股市的交叉互相关性.利用最大重复离散小波变换对沪深股市的交叉互相关性做小波分析,把互相关函数分解在不同的尺度上,以便更清晰地识别沪深股市高频收益序列的互相关性.经过尺度分解后,互相关序列的小波方差之和等于原序列的方差,原高频互相关序列高峰、厚尾的特性不再显著,并趋向于正态分布.

关键词: 高频, 小波变换, 互相关, 小波方差

Abstract: Cross-correlation of high-frequency time series between Shanghai Stock Market and Shenzhen Stock Market is studied.In order to identify details of cross-correlation more distinctly,cross-correlation is analyzed by MODWT(maximal overlap discrete wavelet transform)on various scales.The sum of wavelet variance after decomposition is equal to the variance of original time series.Along with decomposition,the characteristic of high kurtosis and thick tail has been decaying;distribution of new time series is tending to normal distribution.

Key words: high-frequency, wavelet transform, cross-correlation, wavelet variance

中图分类号: