Variance-Stable R-Estimators.
Abstract
By means of the concept of change-of variance function we investigate the stability properties of the asymptotic variance of R-estimators. This allows us to construct the optimal V-robust R-estimator that minimizes the asymptotic variance at the model, under the side condition of a bounded change-of variance function. Finally, we discuss the connection between this function and an influence function for two-sample rank tests introduced by Eplett (1980). (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1984
- Accession Number
- ADA143172
Entities
People
- E. Ronchetti
- J. H. Yen
Organizations
- Princeton University