ANALYSIS OF MARGINAL COUNT FAILURE DATA WITH DISCARDING INFORMATION BASED ON LFP MODEL

摘要

This paper discusses the problem of parametric estimation of failure time distribution from marginal count failure data for product populations where failures are observed only from the units with defects before they have been discarded and no failure occurs in the nondefective units. Assuming that failure times follow a Weibull distribution, we propose a likelihood-based method for estimating the parameters of failure time distribution and p, the proportion of the defective products in the population. The estimation algorithm is described through an application to an actual data set of this method.