ABSTRACT
Proportional hazards is a property of survival model that is fundamental for the development of non parametric regression models. After defining this property, the logrank test is introduced for comparing afinite number of survival distributions. It is in fact a special case of Cox regression, which is treated next.That includes introducing a method for estimating the baseline cumulative hazards and survival functions.The proportional hazards property is generalized to discrete-time survival modeling, and an analogue to Coxregression, mased on maximum likelihood, is introduced. All comes with real-life data examples.
