Closed hfrick closed 9 months ago
Chatted with Max: yes to covariates for the reverse KM in general but because those could be their own set of covariates (different from those on the even time distribution) adding them is complex. So, until we do that, we might as well remove/deprecate this argument.
This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.
The
predict()
method forcensoring_model_reverse_km
models has anew_data
argument which is passed on to the predict method for prodlim objects. The definition for this arg from prodlim:However, the right hand side for
censoring_model_reverse_km
models is always just1
, no covariates: https://github.com/tidymodels/parsnip/blob/2a8f92e1a12638c8b2dfe8e47fb56923c752828a/R/survival-censoring-model.R#L27Because of that, the argument doesn't do anything, you can even feed it a totally unrelated dataset and it won't complain. Do we anticipate adding covariates to that reverse KM anytime soon/ever? If not, can we deprecate that arg to reduce complexity we are not even making use of?
Created on 2023-05-11 with reprex v2.0.2