The Bias in the Roots of Sample Covariance Matrices.
Abstract
Many statistical techniques involve the use of the roots of sample covariance matrices, and in general, the roots of Wishart matrices. Any bias in these roots should be taken into account when these methods are employed. It is demonstrated analytically that for certain cases, the roots are biased and the exact bias is computed for some situations. A Monte Carlo simulation is performed for more general cases and the bias is again apparent. The effects of the bias in the study of principal components and in the estimation of the inverse of Wishart matrices are discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1976
- Accession Number
- ADA023490
Entities
People
- Lyman L. Mcdonald
- Robert K. Smidt
Organizations
- University of Wyoming