摘要
股票市场超高频数据具有交易间隔随机性的特点,传统的皮尔逊相关性度量不能够直接使用原始交易数据,需要通过插值得到均匀、同步的抽样序列.傅里叶分析法不需要对原始数据进行插值,能更精确地度量时间序列的相关性.将傅里叶分析法用于我国金融市场股票收益率的相关性分析中,对皮尔逊相关分析和傅里叶分析法的度量效果进行了比较.
A fundamental property of financial high-frequency data is that the sampled data can occur at random, standard correlation measures can not be directly applied to the raw time series, which have to be homogenized and synchronized by interpolation. The method based on Fourier analysis is an alternative method which can be applied directly to the actual time series. In this paper, Fourier analysis method is used to measure the correlation of stocks, the efficiency of Fourier method is given by empirical study.
出处
《数学的实践与认识》
CSCD
北大核心
2010年第7期63-67,共5页
Mathematics in Practice and Theory
基金
北京市教委优秀人才培养基金(SM200710005001)
关键词
超高频数据
傅里叶分析法
相关性度量
ultra high-frequency data
fourier analysis method
pearson correlation coefficient
correlation measure
