Inference on the Ranks of the Canonical Correlation Matrices for Elliptically Symmetric Populations.
Abstract
In this paper, the authors considered the likelihood ratio tests and some other tests for the ranks of the canonical correlation matrices when the underlying distributions are real and complex elliptically symmetric distributions. Also, asymptotic joint distributions of the eigenvalues of the sample canonical correlation matrices are derived under the assumptions mentioned above regarding the underlying distributions. Finally, applications of tests for the rank of the complex canonical correlation matrix in the area of time series in the frequency domain are discussed. Originator-supplied keywords: Asymptotic distributions, Complex distributions, Canonical correlations; Elliptical distribution; Time series.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1985
- Accession Number
- ADA158268
Entities
People
- Jian Lin
- Liqiang Wang
- Paruchuri R. Krishnaiah
Organizations
- University of Pittsburgh