Spline nonparametric quasi-likelihood regression within the frame of the accelerated failure time model

The accelerated failure time model provides direct physical interpretation for right censored data. However, the homogeneity of variance assumption of the log transformed data does not always hold. In this paper, we propose using a generalized linear model for right censored data in which we relax the homogeneity assumption. A new semiparametric analysis method is proposed for this model. The method uses nonparametric quasi-likelihood in which the variance function is estimated by polynomial spline regression. This is based on squared residuals from an initial model fit. The rate of convergence of the nonparametric variance function estimator is derived. It is shown that the regression coefficient estimators are asymptotically normally distributed. Simulations show that for finite samples the proposed nonparametric quasi-likelihood method performs well. The new method is illustrated with one dataset. Published by Elsevier B.V.