Dear Prof. Daniele Porcheddu The exponential, Weibull, Gamma, Log-likelihood, Pareto, Logistic distribution etc. These are the most known probability distribution for modelling survival times. Which of these distribution fits your data best, you can use it. For specific cases (such as censoring, finite ranges (e.g. 0
If the survival rate is constant, the survival plot would be a straight line (slope = survival rate = a constant) - right? So, the non-parametric Kaplan-Meier model will work (as in "Keep It Simple").