The Application of Invariance to Unbiased Estimation,
Abstract
A relatively easy method is discussed for constructing minimum variance unbiased estimates of nonstandard parametric functions under a variety of statistical model assumptions. The most important assumption is that the family of given distributions be invariant under a group of transformations. Estimation of the reliability function of an aircraft is an important application, as is estimation of moments of transformed data. The results are applied to unbiased estimation in the multivariate normal distribution and to a problem involving U-statistics. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1970
- Accession Number
- AD0714067
Entities
People
- Carl N. Morris
- Morris L. Eaton
Organizations
- RAND Corporation