Online Program Home
My Program

Abstract Details

Activity Number: 193
Type: Contributed
Date/Time: Monday, August 1, 2016 : 10:30 AM to 12:20 PM
Sponsor: Section on Physical and Engineering Sciences
Abstract #318747
Title: Maximum Likelihood and Least Squares Estimation Comparison for the Three-Parameter Weibull Distribution: Case Study of Statistical Software
Author(s): William Harper* and Thomas R. James
Companies: Otterbein University and Otterbein University
Keywords: Weibull ; software ; reliability ; maximum likelihood ; least squares
Abstract:

Both Maximum Likelihood and Least Squares estimation of the 2-parameter Weibull distribution are straightforward; however, there are concerns for their estimation of the 3-parameter Weibull. The third parameter for the 3-parameter Weibull distribution shifts the origin from 0 to some generally positive value sometimes called the location, threshold, or minimum life. The different methods used by the packages sometimes result in fairly major differences in the estimated parameters between the statistical packages. This has been covered previously using modern statistical software separately for maximum likelihood and least squares (maximum likelihood, 2011: Harper, James, Eschenbach; least squares 2015: Harper, James). This publication directly compares the maximum likelihood and least squares results. The findings may have implications for those needing to estimate or apply the results of a 3-parameter Weibull distribution that is used frequently in practice. The results are analyzed based on an experimental design using pseudo-random Weibull data sets.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

 
 
Copyright © American Statistical Association