Journal of University of Chinese Academy of Sciences >
2012 >Issue 2: 145 - 153
Estimations of the improper linear regression models with complex-valued data
Received date: 2011-03-11
Revised date: 2011-04-02
Online published: 2012-03-15
Supported by
Supported by National Natural Science Foundation of China (10801133) and President Foundation of GUCAS
A complex-valued random variable is improper if its complementary covariance is not 0, otherwise it is proper. In this paper, we consider the linear regression models whose errors are independent and identically distributed following an improper complex Gaussian distribution. Such a model may be encountered in signal processing. We use the maximum likelihood method and the two-stage least squares method to estimate the regression coefficients. Differences between these two methods and the complex version of least squares estimates are investigated by simulations. A wind data set is used to illustrate the proposed approach.
Key words： complex-valued data; improper; complex Gaussian distribution
YAN Chang-Jun , ZHANG San-Guo , NING Wei . Estimations of the improper linear regression models with complex-valued data[J]. Journal of University of Chinese Academy of Sciences, 2012 , (2) : 145 -153 . DOI: 10.7523/j.issn.2095-6134.2012.2.001
[1] Goodman J W. Statistical optics[M]. Wiley, 1985:1-5.
[2] Schreier P J, Scharf L L. Statistical signal processing of complex-valued data: the theory of improper and noncircular signals[M]. London: Cambridge University Press, 2009:22-48.
[3] Richards M A. Fundamentals of radar signal processing[M]. McGraw-Hill, 2005:1-52.
[4] Gu X Q, Kang H W, Jiang J M. Monthly temperature forecasts by using a complex autoregressive model[J]. Journal of Applied Meteorological Science, 2005, 18(4):435-440(in Chinese). 谷湘潜,康红文,江剑民. 用复数自回归模式预报月平均气温[J].应用气象学报,2005,18(4):435-440.
[5] Picinbono B. Second-order complex random vectors and Normal distribution[J]. IEEE Transaction on Signal Processing, 1996, 44 (10):2637-2640.
[6] Schreier P J, Scharf L L, Hanssen A. A generalized likelihood ratio test for impropriety of complex signals[J]. IEEE Signal Procession Letter, 2006, 13(7):433-436.
[7] Van den Bos A. The multivariate complex normal distibution-A generalization[J]. IEEE Transaction on Information Theory, 1995, 41(2):537-539.
[8] Picinbono B. On circularity[J]. IEEE Transaction on Signal Processing, 1994,42(12):3473-3482.
[9] Miller K S. Complex linear least squares[J]. Siam Review, 1973, 15(4):706-726.
[10] Picinbono B, Chevalie P. Widely linear estimation with complex data[J]. IEEE Transaction on Signal Processing, 1995, 43(8):2030-2033.
[11] Navarro-Moreno J. ARMA prediction of widely linear systems by using the innovations algorithm[J]. IEEE Transaction on Signal Processing, 2008, 56(7):3061-3068.
[12] Navarro-Moreno J, Moreno-Kayser J, Fernández-Alcalá R M, et al. Widely linear estimation algorithms for second-order stationary signals[J]. IEEE Transaction on Signal Processing, 2009, 57(12):4930-4935.
[13] Mandic D P, Goh S L. Complex valued nonlinear adaptive filters: noncircularity, widely linear and neural models[M]. John Wiley & Sons, 2009:13-31.
[14] Andersen H H, Hojbjerre M, Sorensen D, et al. Linear and graphical models for the multivariate complex normal distribution[M]. Springer-Verlag, 1995:1-7.
[15] Hooper P M. Iterative weighted least squares estimation in heteroscedastic linear models[J]. Journal of the American Statistical Association, 1993, 88: 179-184.
[16] Zhao Y, Hu Y J, Huang W. Effect of wind farm fluctuation power on grid voltage[J]. Jilin Electric Power, 2007, 35(2):22-24(in Chinese). 赵勇,胡雅娟,黄巍. 风电场功率波动对电网电压的影响[J]. 吉林电力,2007,35(2):22-24.
/
| 〈 |
|
〉 |