The Ubiquitous Role of f'/f in Efficiency Robust Estimation of Location.

Abstract

This paper is primarily expository in nature and focuses on the all pervasive importance of f'/f in efficient estimation of location, with primary emphasis on the role of f'/f in robust estimation. Connections between M estimators (maximum likelihood-like), R (rank) estimators and L estimators (linear combinations of order statistics) are discussed and an alternative heuristic explanation of f'/f is given showing why it is an intuitively reasonable quantity on which to base estimation. The asymptotic relative efficiency of each class of estimators is shown to be the square of a correlation coefficient related to f'/f and reasons are given as to why R estimators might often prove to have superior robustness properties relative to L and M estimators. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1980
Accession Number
ADA093563

Entities

People

  • Brian L. Joiner
  • David L. Hall

Organizations

  • University of Wisconsin–Madison

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Coefficients
  • Data Science
  • Distribution Functions
  • Estimators
  • Gaussian Distributions
  • Information Science
  • Mathematics
  • Maximum Likelihood Estimation
  • Normal Distribution
  • Order Statistics
  • Probability
  • Statistical Algorithms
  • Statistical Analysis
  • Statistics
  • United States
  • Universities
  • Wisconsin

Fields of Study

  • Mathematics

Readers

  • Regression Analysis.
  • Robotics and Automation.
  • Theoretical Analysis.