Modified Kernel Smoothers for the Right-Censored Partially Linear Models

Survival analysis poses challenges when dealing with right-censored data, characterized by an incomplete structure. Classical estimation methods in the data modeling context face limitations in the presence of right-censored response variables, necessitating a solution to the censorship problem before modeling. This study addresses the censorship challenge using synthetic data transformation, Kaplan-Meier weights, and Buckley-James transformation. In methodology, we focus on modifying the semiparametric kernel smoothers according to the censorship solutions and introducing alternative estimators for the right-censored partially linear model. To present comparative performances of the modified kernel estimators, a simulation study is carried out.