Cameron Davidson-Pilon
Cameron Davidson-Pilon
Hi @MetzgerSK - thanks for the (very) detailed report. I'll look into this soon. Apologies that this is occurring. I'm relieved that a previous-me did write [tests for this function](https://github.com/CamDavidsonPilon/lifelines/blob/71f4e4d659870058c33572e7e582654f96c1fcaa/lifelines/tests/test_statistics.py#L277-L492),...
I haven't yet dug into this, but my suspicion is that the results are due to how ties are handled. `rossi` has lots of ties, whereas the testing dataset I...
Hi @louisazhou, this looks interesting, but I'm worried that this is a _documentation problem_ instead. When using `conditional_after`, the timeline is moved up to _start_ at the value of `conditional_after`....
We can rewrite your equation as: So really the conditional cumulative hazard rate is H(t2 | t_1) = H(t_2) - H(t_1). Do you agree? By definition, H(t) is an non-decreasing...
If you disagree, please continue asking questions! This is useful for me, and there may be a subtle bug here.
I'm happy it's resolved, but I'm going to reopen this issue as a reminder to make the docs more clear
Hi @aaronsmith1234, no opposition here! You'll find _some_ type hinting scattered throughout already, but always happy to see more!
Agree! This would be a nice addition, thanks for the suggestion!
Hi @pisonir - it's on the roadmap! First is to implement confidence intervals for the `predict_` methods, and then extending them to `plot_partial_effects_on_outcome`.
Also https://lifelines.readthedocs.io/en/latest/Survival%20Regression.html#cox-s-proportional-hazard-model