Abstract
Analysis of product lifetime based only on warranty data is not informative. Some papers have proposed several interesting estimators where supplementary information is gathered from other sources. The monthly(or weekly, etc.)unit sales are regarded as necessary information in almost all the existing works. However, for many products it is not easy to obtain monthly unit sales information from many of the wholesalers and retailers around the world. This paper deals with the estimation problem from warranty claim data, when monthly unit sales information is not available, but the total sales number can be determined from other sources. We derive nonparametric maximum likelihood estimators(MLEs)and give their asymptotic properties. The results of our simulations confirm that the proposed MLEs can give accurate estimates of lifetime distribution functions.
