Goodness of Fit Tests for the Weibull and the Extreme Value Distribution with Estimated Parameters.
Abstract
In this paper we consider the problem of testing the null hypothesis that a given random sample belongs to a Weibull of an extreme value distribution with unknown parameters. The test statistics are those based on the empirical distribution function, and tables of critical values are provided. The asymptotic points have been obtained by a pooling of two methods. In the first method the percentage points for finite n are plotted and extrapolated to infinity. In the second method, the appropriate asymptotic process is simulated and its percentiles, which give the critical points, thus estimated. Some difficulties in simulating the asymptotic process are discussed, and a comparison between two methods is discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 20, 1979
- Accession Number
- ADA096981
Entities
People
- Mahesh Chandra
- Michael A. Stephens
- Nozer Singpurwalla
Organizations
- George Washington University