作者: C. R. RAO
DOI: 10.1093/BIOMET/52.3-4.447
关键词: Multivariate random variable 、 Centering matrix 、 Square matrix 、 Mathematical optimization 、 Linear least squares 、 Covariance matrix 、 Mathematics 、 Hilbert matrix 、 Applied mathematics 、 Pascal matrix 、 Non-linear least squares
摘要: In an earlier paper (Rao, 1959), the author discussed method of least squares when observations are dependent and dispersion matrix is unknown but independent estimate available. The was, however, considered as arbitrary positive definite matrix. present we shall consider a class problems where has known structure discuss appropriate statistical methods. More specifically results from considering parameters in well-known Gauss-Markoff linear model random variables. Let Y be vector variable with