A Modified Kolmogorov-Smirnov Test for the Gamma and Weibull Distribution with Unknown Location and Scale Parameters.
AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING
Pagination or Media Count:
The Kolmogorov-Smirnov tables for the Weibull and Gamma distribution were generated in this thesis. These tables are used when testing whether a set of observations are from a Weibull Gamma distribution in which the location and scale parameters must be estimated from the sample. A power investigation of the test against some selected distributions using Monte Carlo techniques was conducted. This procedure shows that the test is reasonably powerful for a number of alternative distributions. A relationship between the critical values and the shape parameters was investigated for the Weibull and Gamma distributions. No apparent relationship was found for the Gamma distribution. In contrast, an approximate log-linear relationship was found for the Weibull when the shape parameter is between one and four. Author
- Statistics and Probability