摘要
以高频数据为载体,研究它的日内周期性和日间趋势性;利用小波分析中的多分辨分析方法对短期周期性和长期趋势性进行分离.对多分辨分析后的效果进行检验,发现对日内周期性滤波的拟合度比FFF方法高.多分辨分析方法能更好地识别高频部分的细节,有助于对金融市场微观结构的研究.
In this paper intraday periodicity and interday trend of stock market＇s high-frequency data is studied. With MRA （Multiresolution Analysis）, short-term periodicity and long-term trend is taken apart. The empirical results demonstrate that the goodness of fit of MRA is better than that of FFF （Flexible Fourier Function） in filtering intraday periodicity. Furthermore, MRA can identify the details of high-frequency data more easily and make for research of financial market＇s microstructure.
出处
《河北工业大学学报》
CAS
2005年第4期38-43,共6页
Journal of Hebei University of Technology
基金
国家自然科学基金资助项目(70471050)
关键词
日历效应
小波分析
多分辨分析
能量谱
calendar effect
wavelet analysis
multiresolution analysis
power spectrum
