Authors: Mostafa Karimi and Noor Akma Ibrahim
Abstract: Estimating the covariance matrix of the rank estimators in the accelerated failure time model is complicated in the presence of interval censored data. The main difficulty with the existing estimation methods is that they involve nonparametric approximation of the density function of error terms. In this paper a resampling method for approximating the covariance matrix of the rank estimators is developed which is established on the basis of interval censored data. The proposed resampling scheme is assessed through a real example.
Keywords: component; accelerated failure time model; Gehan statistic; weighted log-rank statistic; covariance matrix; interval censored data