Publication at Faculty of Mathematics and Physics |
2015
Abstract
When observing independent devices which are subject to degradation, we want to describe the influence of repairs and preventive maintenance actions on the time to failure distribution with the help of suitable regression models. Commonly used models, as the Cox proportional hazards and the accelerated failure time model assume, that the covariates influence a certain baseline hazard function, which must be either parametrized or estimated nonparametrically.
In this work we focus on methods how to estimate and test hypotheses about the shape of the baseline hazard and we show their applications.