I'm just starting to learn about Survival Analysis and I understand the theoretical proofs but I'm still unclear about how to do some of the estimations practically.


The primary complication in the analysis of survival data is censoring, i.e., the fact that you do not observe the event of interest for some subjects in your study. Many statistical procedures do not work with censored data, such as the empirical cumulative distribution function.

To obtain valid estimates you will need to use statistical techniques that do account for it. If you have right censored data, you can obtain a non-parametric estimate of the cumulative hazard function using the Nelson-Aalen estimator. Another option is to first obtain an estimate of the survival function using the Kaplan-Meier estimator and then calculate the minus logarithm of it to obtain the cumulative hazard function.

  • $\begingroup$ I see now appreciate the answer ! In the case I'm trying to model default probabilities of a sample of firms, can I use the procedure you suggested to obtain the Cumulative Hazard and then compute the Instantaneous Hazard to finally get Default Probabilities? Because I've read some literature that said that the Instantaneous Hazard isn't actually a probability. Thanks. $\endgroup$ – Metrician Aug 3 at 13:02
  • $\begingroup$ To put it in a less convoluted way: is there a way, using the approach you just suggested, to estimate Hazard Rates from the Cumulative Hazard? $\endgroup$ – Metrician Aug 3 at 13:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.