Robustness properties for a simple class of rank estimates
Thomas P. Hettmansperger, Jessica Utts
Abstract
Thomas P. Hettmansperger, Jessica Utts
Abstract
Robustness properties of a family of rank estimates to compete with trimmed means and other robust estimates for the one sample location problem are investigated. In particular, the influence curve and breakdown point are developed, as well as their finite sample equivalents, the sensitivity curve and tolerance. The estimates are formulated from a one sample rank test rather than the customary two sample rank test approach. In addition, a functional is implicitly defined for the asymptotic version of the estimate. Computational problems are considered and a simple iterative procedure for finding the estimate is given.
OpenAlex reports 9 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Robustness properties of a family of rank estimates to compete with trimmed means and other robust estimates for the one sample location problem are investigated. In particular, the influence curve and breakdown point are developed, as well as their finite sample equivalents, the sensitivity curve and tolerance. The estimates are formulated from a one sample rank test rather than the customary two sample rank test approach. In addition, a functional is implicitly defined for the asymptotic version of the estimate. Computational problems are considered and a simple iterative procedure for finding the estimate is given.
Key concepts: Robustness (evolution), Mathematics, Rank (graph theory), Applied mathematics, Simple (philosophy), Sample (material), Mathematical optimization, Combinatorics