Recursive Estimation Procedures for Missing-Data Problems.

Abstract

Titterington in a previous paper proposed recursive methods for dealing with incomplete data. This present paper concentrates on versions of these for multiparameter problems involving missing data. Theorems are outlined from which asymptotic properties of the recursive procedures can be established and versions of the recursions are written down for problems in which the missing data are missing at random. After illustration with exponential family models, the case of multivariate Normal data is considered in detail. Numerical comparisons of the various methods are obtained using bivariate Normal data. Whereas the previous paper discussed incomplete data in general, the present one restricts attention to the problem of missing values. Typically, each experimental unit should have records of the values of several characteristics associated with it. Statistical analysis is made difficult if one or more of those values are missing on some units. To combat the heavy analysis required for a 'proper' analysis of the data, comparatively simple recursive procedures are outlined in which the data are incorporated sequentially into the estimation scheme. Some comments are made about theoretical properties and special emphasis is laid on the case of data from multivariate Normal distributions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1982
Accession Number
ADA124358

Entities

People

  • D. M. Titterington
  • J-m. Jiang

Organizations

  • University of Wisconsin–Madison

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Computational Science
  • Data Science
  • Data Sets
  • Estimators
  • Information Science
  • Mathematics
  • Maximum Likelihood Estimation
  • Normal Distribution
  • Observation
  • Probability
  • Statistical Algorithms
  • Statistical Analysis
  • Statistics
  • Surveys
  • United States
  • Universities

Fields of Study

  • Mathematics

Readers

  • Statistical inference.
  • Systems Analysis and Design