Generalized Logit-Based Proportional Hazards Models and Their Applications in Survival and Reliability Analyses

We introduce a flexible family of generalized logit-based regression models for survival and reliability analyses. We present its parametric as well as its semiparametric versions. The method of maximum likelihood and the partial likelihood approach are applied to estimate the parameters of the parametric and semiparametric models, respectively. This new family of models is illustrated with male laryngeal cancer data and compared with Cox regression.