On Rules Based on Sample Medians for Selection of the Largest Location Parameter.
Abstract
The problem of selection of a subset containing the largest of several given location parameters is considered, and selection rules based on sample medians are proposed and investigated. Selection of the largest normal mean is considered in detail and some new results for double exponential populations are also obtained. Numerical comparisons in the case of equally spaced normal means are made between the medians procedure and the means procedure. The asymptotic relative efficiency (ARE) of the medians procedure relative to the means procedure is also computed, assuming that the normal means are in a slippage configuration. The means procedure is found to be much better than the medians procedure in the sense of ARE. However, if the normal populations are highly contaminated, the proposed rule based on sample medians is superior to the means procedure.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1977
- Accession Number
- ADA047948
Entities
People
- Ashok K. Singh
- Shanti Gupta
Organizations
- Purdue University