Nearly Best Linear Unbiased Conditional Estimators of the Location and Scale Parameters of the Normal Distribution by Use of Order Statistics.
Abstract
Lagrange multipliers are used to develop conditional estimators, based upon order statistics, for the location and scale parameters of continuous distributions. Tabulated coefficients for nearly best, linear, unbiased, conditional estimators are presented. These coefficients are designed for use with uncensored, singly censored, or doubly censored samples of size 2 through 40 drawn from a population which may be approximated by the normal distribution. The efficiency of these estimators (compared to the best, linear, unbiased, conditional estimators) is above 98 per cent for all sample sizes. The tabulated coefficients may also be used to estimate either or both parameters unconditionally. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1969
- Accession Number
- AD0776259
Entities
People
- James L. Socolofsky
Organizations
- Air Force Institute of Technology