Test of independence between tow sets of variates (Q3223686)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3890509
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Test of independence between tow sets of variates |
scientific article; zbMATH DE number 3890509 |
Statements
Test of independence between tow sets of variates (English)
0 references
1984
0 references
asymptotic expansions
0 references
likelihood ratio criterion
0 references
testing independence
0 references
non-central beta distributions
0 references
canonical correlation coefficients
0 references
0 references
0 references
0.9120762944221495
0 references
0.9120762944221495
0 references
0.8036040663719177
0 references
0.8014377355575562
0 references
In this paper asymptotic expansions of the null as well as non-null distributions of the likelihood ratio criterion for testing independence between two sets of variates are obtained. In the null case for \(p=1\) and \(p=2\), the expansion reduces to the exact distribution. In the non-null case, the expansion is given in terms of non-central beta distributions and for the case when the population canonical correlation coefficients are small. In more detail, let us partition the covariance matrix \(\Sigma\) and S (\(=\sum^{n}_{\alpha =1}(x_{\alpha}-\bar x)(x_{\alpha}-\bar x)\)) into p and q rows and columns as \[ \Sigma =\left( \begin{matrix} \Sigma_{11}\quad \Sigma_{12}\\ \Sigma_{12}\quad \Sigma_{22}\end{matrix} \right),\quad S=\left( \begin{matrix} S_{11}\quad S_{12}\\ S_{12}\quad S_{22}\end{matrix} \right). \] Then the likelihood ratio criterion for testing \(H:\Sigma_{12}=0\) against all alternatives \(K: \Sigma\) \({}_{12}\neq 0\) is given by \(\lambda =(| S| /| S_{11}| \cdot | S_{12}|)^{N/2}.\) Assume \(w=\lambda^{2/N}\), and \(v=w^{1/r}\) where r is a convergence factor. Both null and non-null distributions of w or v are obtained.
0 references