Asymptotic distributions of two test statistics for testing independence with missing data

From MaRDI portal





Analysis of multivariate data with missing observations is the topic of the paper. It considers one of the many problems discussed by \textit{M. L. Eaton} and \textit{T. Kariya} [Ann. Stat. 11, 654-665 (1983; Zbl 0524.62051)]. It is desired to test the independence of two sets of variates of dimensions \(p_ 1\) and \(p_ 2\) where \(p_ 1+p_ 2=p\), when data based on p-coordinates are such that you have N p-dimensional observation vectors and \(M_ i\) \(p_ i\)-dimensional observation vectors, \(i=1,2\). All \(N+M_ 1+M_ 2\) observations are assumed to be independent. Thus there are N complete observations, \(M_ 1\) extra observations on first \(p_ 1\) coordinates and \(M_ 2\) extra observations on last \(p_ 2\) coordinates and \(N>0\), \(M_ 1>0\), \(M_ 2>0\).











This page was built for publication: Asymptotic distributions of two test statistics for testing independence with missing data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3736737)