On the distribution of a statistic in multivariate inverse regression analysis (Q2266546)

Let \(y=(y_ 1,...,y_ p)'\) be random, \(x=(x_ 1,...,x_ q)'\) be fixed vectors and \(y=a+B'x+e=\theta '[1,x]'+e\) where \(\theta '=[a\), B'] is the \(p\times (1+q)\) matrix of unknown parameters and e is an error vector having a multivariate normal distribution \(N_ p(0,\Sigma)\). Further, suppose that N informative observations on y and x have been given. In this paper, the following problems are discussed: (a) Derivation of an asymptotic expansion for the distribution function of the statistic up to the order \(N^{-2}\). (b) The distribution problem in the situation where \(\theta\) is known and \(\Sigma\) is unknown. (c) The distribution problem when \(\theta\) and \(\Sigma\) are unknown.