Open
Description
Hello :)
First of all thanks for bringing MTLR on R !
I was wondering if you have some recommendations about the binning of the time interval, and consequently about the time points to use. Would it make sense, for instance, to use the unique observed times of a training dataset (censored or not) as time points ?
If I understood well, extraction of survival probabilities at some requested time points are obtained with a spline. Would it benefit from the use of a large number of time points in the training process ?
Thank you in advance !
Have a nice day !
Metadata
Metadata
Assignees
Labels
No labels