Some overly strict code in predict.flexsurvreg() did not allow predictions to be made at time zero, which should indeed be allowed as per summary.flexsurvreg(). This has now been fixed.
The nested list-columns which are returned when requesting multiple predictions per subject carried along some additional classes that were not needed. I have now used a safer function for nesting and this issue is resolved.
While I was at it, I worked on cleaning up some of the error messages to be cleaner and more accurate.
Please let me know if you would like to see any changes.
Hi @chjackson,
As noted by @hfrick over in https://github.com/tidymodels/censored/issues/152#issuecomment-1076215197, there were two issues in
predict.flexsurvreg()
that I have addressed with this PR:predict.flexsurvreg()
did not allow predictions to be made at time zero, which should indeed be allowed as persummary.flexsurvreg()
. This has now been fixed.While I was at it, I worked on cleaning up some of the error messages to be cleaner and more accurate.
Please let me know if you would like to see any changes.