Easy Bayes Estimation for Rasch-Type Models.

Abstract

A Bayes estimation procedures is introduced that allows the nature and strength of prior beliefs to be easily specified and posterior models to be estimated with no more difficulty than maximum likelihood estimation. The procedure is based on constructing posterior distributions that are formally identical to likelihoods, but are constructed partly from sample data and partly from artificial data reflecting prior information. Improvements in performance of modal bayes procedures relative to maximum likelihood likelihood estimation procedures are illustrated for Rash-type models. Improvements range from modest to dramatic, depending on the model and the number of items being considered.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 04, 1987
Accession Number
ADA193628

Entities

People

  • James E. Laughlin
  • Kai F. Yu
  • Robert J. Jannarone

Organizations

  • University of South Carolina

Tags

Communities of Interest

  • Autonomy
  • C4I
  • Human Systems

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Artificial Intelligence Software
  • Classification
  • Computational Science
  • Identification
  • Machine Learning
  • Markov Models
  • Maximum Likelihood Estimation
  • Military Research
  • New York
  • Probability
  • Security
  • Social Sciences
  • South Carolina
  • Statistics
  • United States
  • Universities

Fields of Study

  • Mathematics

Readers

  • Artificial Intelligence
  • Statistical inference.
  • Systems Analysis and Design