Open gowerc opened 6 months ago
For reference this is pseudo already implemented via the sampleSubjects.SimSurvival
method which could have sub-class specific implementations. Only problem is that sampleObservations.SimSurvival
appears to be hard coded to look for the cov_cat
and cov_cont
variables which would need to be fixed. I'm not sure where else these variables are hard coded either. Might need some thinking about detaching this definition from the baseline-hazard though as at the moment you'd have to redefine all your survival distributions if you change the method (maybe this is fine?)
Currently the survival simulation function is hard coded to only have 2 linear covariates
cov_cat
andcov_cont
. Would be good to extend this somehow to allow for a more flexible specification from theSimSurvial
objects.