Semiparametric Additive Accelerated Life Models (Q4956040)

The topic of time transformation functions of the accelerated life test is of great importance. Among all the time transformation functions the most important seems to be the transfer functional which allows one to obtain quantiles of the same level for the reliability of items submitted to different stresses. \textit{V. Bagdonavičius} and \textit{M. S. Nikulin} [``Semiparametric models in accelerated life testing'' (1995; Zbl 0839.62092) and Biometrika 84, No. 2, 365-378 (1997)] gave several classes of models based on properties of the transfer functionals. They showed that their models contain practically all the known models of accelerated life testing and even semiparametric models of survival analysis like the well-known Cox proportional hazards model [see \textit{D. R. Cox}, J. R. Stat. Soc., Ser. B. 34, 187-220 (1972; Zbl 0243.62041)], the additive hazards model of \textit{D. R. Cox} and \textit{D. Oakes} [``Analysis of survival data'' (London, 1984)], etc. Moreover, they proposed different methods of inference in their models (with right censored data) using counting processes technique. The asymptotic properties of estimators obtained for the semiparametric additive accelerated life model, proposed by \textit{V. Bagdonavičius} and \textit{M. S. Nikulin} (loc. cit), are investigated. This model generalizes the well-known additive hazards model of survival analysis and is close to the general transformation model. Asymptotic properties of the estimator of the regression parameter and the estimator of the reliability function are given in the case of right censoring for discretized data, and a numerical example illustrates these results (using Monte Carlo studies).