Understanding and Addressing the Unbounded “Likelihood” Problem

Thumbnail Image
Date
2013-09-01
Authors
Liu, Shiyao
Wu, Huaiqing
Meeker, William
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Authors
Person
Meeker, William
Distinguished Professor
Research Projects
Organizational Units
Organizational Unit
Statistics
As leaders in statistical research, collaboration, and education, the Department of Statistics at Iowa State University offers students an education like no other. We are committed to our mission of developing and applying statistical methods, and proud of our award-winning students and faculty.
Journal Issue
Is Version Of
Versions
Series
Department
Statistics
Abstract

The joint probability density function, evaluated at the observed data, is commonly used as the likelihood function to compute maximum likelihood estimates. For some models, however, there exist paths in the parameter space along which this density-approximation likelihood goes to infinity and maximum likelihood estimation breaks down. In applications, all observed data are discrete due to the round-off or grouping error of measurements. The “correct likelihood” based on interval censoring can eliminate the problem of an unbounded likelihood. This paper categorizes the models leading to unbounded likelihoods into three groups and illustrates the density breakdown with specific examples. We also study the effect of the round-off error on estimation, and give a sufficient condition for the joint density to provide the same maximum likelihood estimate as the correct likelihood, as the round-off error goes to zero.

Comments

This preprint was published as Shiyao LIu, Huaiqing Wu & William Q. Meeker, "Understanding and Addressing the Unbounded "Likelihood" Problem", The American Statistician (2015): doi: 10.1080/00031305.2014.1003968.

Description
Keywords
Citation
DOI
Source
Subject Categories
Copyright
Collections