Random Forest minimal depth filter returns all NA's

[annette987]

The random forest minimal depth filter is returning all NA values. This is because the vector of features returned by the filter is not a named vector and generateFilterValuesData expects a named vector. I can submit a PR to fix this and it will explain the solution.

library(survival)
library(mlr)
#> Loading required package: ParamHelpers
#> 'mlr' is in maintenance mode since July 2019. Future development
#> efforts will go into its successor 'mlr3' (<https://mlr3.mlr-org.com>).

set.seed(24601)
data(veteran)
mas.task <- makeSurvTask(id = "VET", data = veteran, target = c("time", "status"))
outer = makeResampleDesc("CV", iters = 2, stratify = TRUE)

cox.filt.mindepth.lrn = makeFilterWrapper(
    makeLearner(cl = "surv.coxph", id = "cox.filt.mindepth", predict.type = "response"), 
    fw.method = "randomForestSRC_var.select",
    fw.abs = 5,
    more.args = list("randomForestSRC_var.select" = list("metho" = "md", ntree=100, nsplit = 10, nodesize = 3))
  )

learners = list(cox.filt.mindepth.lrn)   
bmr = benchmark(learners, mas.task, outer, show.info = TRUE)
#> Task: VET, Learner: cox.filt.mindepth.filtered
#> Resampling: cross-validation
#> Measures:             cindex
#> Error in (function (task, method = "randomForestSRC_importance", fval = NULL, : Filter method returned all NA values!

^{Created on 2020-01-08 by the reprex package (v0.3.0)}

[pat-s]

A PR including a test for this would be great!
It sounds this only applies to this specific filter method, correct?

[annette987]

Yes that's correct.