Use of partial censored regression models: A survival data application

Abstract

A central aspect in research in several scientific fields, such as engineering, medicine and economy, is the construction of models that adequately represent the data generating process for a temporal variable that gathers information regarding reliability, survival or duration. The present work was based in a methodology that extends the traditional regression models used by Cox and Stute in survival analysis, including in their specification the effect of a temporal covariable. The analysis is non-parametrically gathered. The relevant aspect of this work is the form in which the temporal variable effect on the survival variable is treated.
Furthermore, the results from Cox and Stute and the Accelerated Failure Time (AFT) model become particular cases of the partial censored regression model.