Statistical Multiple-Decision Procedures for Some Multivariate Selection Problems
Abstract
The following statistical multiple-decision problems are considered for a multivariate normal distribution with unknown (or partially known) covariance matrix, using the indifference-zone and subset approaches: (a) Selecting the variate with the largest population mean; (b) selecting the variate with the smallest population variance; (c) selecting the subclass of variates with the smallest population generalized variance; (d) selecting the population with the smallest vector coefficient of alienation between two subclasses of variates; (e) selecting the best subclass of predictors for a specified subclass of variates. Small-sample theory is employed in (a) and (b), while large-sample theory is used in (b), (c), (d) and (e). (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1973
- Accession Number
- AD0766469
Entities
People
- Ricardo M. Frischtak
Organizations
- Cornell University