Title :
Research on multiresolution recognition of risk and portolio
Author :
Xu, Qi-fa ; Zhang, Shi-Ying
Author_Institution :
Sch. of Manage., Tianjin Univ., China
Abstract :
This paper presents a new method for multiresolution Betas estimation based on wavelet multiresolution analysis, which decomposes the variance of one time series and the covariance between two time series on a scale by scale basis through maximal overlap discrete wavelet transform. The composition of risk at different scales for high-frequency financial assets is discussed by the method. In accordance with multiresolution character of return and risk, multiresolution portfolio tactics are put forward, which better static portfolio method of Markowitz. The empirical results show that the proposed methods not only capture character of multiresolution, but also put investment risk at different scales down to the minimum level.
Keywords :
covariance analysis; discrete wavelet transforms; investment; risk analysis; time series; Markowitz method; discrete wavelet transform; high-frequency financial assets; investment risk; multiresolution Betas estimation; multiresolution portfolio tactics; multiresolution risk recognition; time series; wavelet multiresolution analysis; Discrete wavelet transforms; Erbium; Forward contracts; Frequency domain analysis; Investments; Multiresolution analysis; Portfolios; Pricing; Time measurement; Wavelet analysis; Betas; Maximal overlap discrete wavelet transform; high-frequency data; multiresolution; portfolio;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527542