Predict censored samples conditional on known duration

[buhrmann]

Hi, it could be useful to include the option to predict remaining survival conditioned on known duration, as the lifelines library does. Alternatively, if that's the correct approach, perhaps something in the docs on how to calculate it oneself, such as dividing the survival curve by S(s) as explained here: https://lifelines.readthedocs.io/en/latest/Survival%20Regression.html#prediction-on-censored-subjects. I imagine that would be valid when using your SurvivalBoost model?

[juAlberge]

Hello!
Thanks for the issue! We'll need to look better at it but something like that can be added!

[Vincent-Maladiere]

Hello @buhrmann, thank you for sharing your thoughts! We're working on implementing your suggestion at #93.

In addition, could you share a little about your use cases in survival analysis / competing risks? In particular, in which context do you need to predict the remaining survival conditioned on known duration? Knowing more about various real-world uses would help us steer the project in the right direction :)