scientific article; zbMATH DE number 3976165

From MaRDI portal
Publication:3740870





[For the entire collection see Zbl 0588.00028.] For a parametric treatment of accelerated life test, this paper develops the basic ingredients of multiple regression analysis under an inverse Gaussian (IG) distribution for the failure time. A plausible regression model is formulated from consideration of the physical aspects of an underlying Wiener process. Asymptotic properties of the maximum likelihood estimator and likelihood ratio statistics are studied, and certain approximations to the sampling distributions are developed by a judicious combination of exact and asymptotic results. The test statistics are related to a decomposition of the total sum of reciprocals into regression, lack of fit, and pure error components. Finally, a graphical method of model checking is introduced by means of constructing the standardized IG residuals.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3740870)