You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We cannot currently pass and store the which.curve parameter from a learner like BART into the PredictionSurv object. Currently, which.curve is just used to calculate the crank in .surv_return(). Therefore, all survival metrics or when combining predictions (via c.PredictionDataSurv) use the median curve (0.5).
TODOs
pass the which.curve in PredictionSurv object somehow
whenever we need to reduce the survival array to a matrix, an Arrdist via the p$distr field to a Matdist or convert with as.Distribution(...), we will extract the which.curve matrix, not the median
combining a mix of survival arrays with matrices is now done by reducing the arrays to the median survival matrices but it should be the which.curve ones
The text was updated successfully, but these errors were encountered:
Description
We cannot currently pass and store the
which.curve
parameter from a learner likeBART
into thePredictionSurv
object. Currently,which.curve
is just used to calculate thecrank
in.surv_return()
. Therefore, all survival metrics or when combining predictions (viac.PredictionDataSurv
) use the median curve (0.5).TODOs
which.curve
inPredictionSurv
object somehowArrdist
via thep$distr
field to aMatdist
or convert withas.Distribution(...)
, we will extract thewhich.curve
matrix, not the medianwhich.curve
onesThe text was updated successfully, but these errors were encountered: