bdwilliamson / vimp

Perform inference on algorithm-agnostic variable importance
http://bdwilliamson.github.io/vimp/
Other
21 stars 8 forks source link

CV not necessary with a single learner (and no hyperparameter permutations) #10

Closed bdwilliamson closed 4 years ago

bdwilliamson commented 4 years ago

If a single learner is entered, run the learner rather than running CV/SuperLearner

bdwilliamson commented 4 years ago

Current fix (with triple-colon) in f57fce8

bdwilliamson commented 4 years ago

Better fix in 8491e18: now, for sp_vim, the user can pass a character vector of learners, in which case SuperLearner is called on the ensemble; or a single function that takes the same arguments as a SuperLearner wrapper (e.g., SL.ranger), in which case the single learner is called with no cross-validation.