Proportional Mean Residual Life Model With Varying Coefficients for Right Censored Data

The mean residual life provides the remaining life expectancy of a subject who has survived to a specific time point. This paper considers a proportional mean residual life model with varying coefficients, which allows one to explore the nonlinear interactions between some covariates and an exposure variable. In a semiparametric setting, we construct local estimating equations to obtain the varying coefficients and establish the asymptotic normality of the proposed estimators. Moreover, the weak convergence property for the local estimator of the baseline mean residual life function is developed. We conduct simulation studies to empirically examine the finite-sample performance of the proposed methods and apply the methodology to a real-life dataset on type 2 diabetic complications.