Robust Estimation Based on Walsh Averages for the General Linear Model.
Abstract
Estimates of the parameters in a linear model are considered based upon the minimization of a dispersion function of the residuals. The dispersion function used depends on Walsh averages of pairs of residuals. Results similar to those arising with signed rank statistics can be obtained as a special case. Trimming and weighting of the Walsh averages can occur with a suitable choice of dispersion function. Asymptotic properties of this type of dispersion function and its derivatives are developed and used to determine the large sample distribution of the estimates. Some discussion appears on the practical application of this methodology. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1983
- Accession Number
- ADA136187
Entities
People
- G. L. Sievers
Organizations
- Western Michigan University