MPIDR Working Paper

Modeling failure (mortality) rate with a change point

Abstract

Simple models for the failure (mortality) rate change point are considered. The relationship with the mean residual lifetime function change point problem is discussed. It is shown that when the change point is random, the observed failure (mortality) rate can be obtained via a specific mixture of lifetime distributions. The shape of the observed failure (mortality) rate is analyzed and the corresponding sim-ple but meaningful example is considered.