Goodness-of-fit tests for reliability modeling

Abstract

In this work we provide goodness-of-fit (GoF) tests which are based on measures of divergence and can be used for assessing the appropriateness of both discrete and continuous distributions. The main focus is on continuous distributions like the exponential, Weibull, Inverse Gaussian, Gamma and lognormal which frequently appear in engineering and reliability. An extensive simulation study shows that the new tests maintain good stability in level and high power across a wider range of distributions and sample sizes than other tests. © Springer-Verlag London Limited 2012.