Support for Survival Forests with time-varying covariates about ranger HOT 6 OPEN

We have currently no plans to implement support for time-varying covariates.
However, it would be great to have that feature! Do you have ideas how to implement this? You are very welcome to contribute the feature to ranger.

from ranger.

It would be a great enhancement. Are u aware of any R package of rsf that deals with time dependent covariates?

from ranger.

No, unfortunately not.

from ranger.

Cox proportional hazards model fitted with the coxph function from survival package can deal with time varying coefficients. See this really neat vignette on the subject: https://cran.r-project.org/web/packages/survival/vignettes/timedep.pdf

from ranger.

It would be a great enhancement. Are u aware of any R package of rsf that deals with time dependent covariates?

R package LTRCforests does implement Random-Survival-Forest with time-varyng covariates.

from ranger.

LTRC does not include multi-core processing.
+1 on needing this feature 6 years later. Minimally I think all would be needed is

1. support for the 'counting process' form of Surv objects [formula = Surv(start, end, status, type = "counting") ~ .]
   a. This currently throws 'Error: Competing risks not supported yet. Use status=1 for events and status=0 for censoring.'
2. optionally, provide arguments to specify groups or id's for subsampling, cross-validation. Subjects should be assigned to the same grouping. The case weights in ranger could be used for obtaining the holdout sample (on user to correctly allocate all records of each subject to only train or hold partitions).

from ranger.