tests/testthat/test_step_select_vip.R

In stevenpawley/recipeselectors: Extra Recipes Steps for Supervised Feature Selection

library(testthat)
library(recipes)
library(tibble)
library(parsnip)
data("iris")

test_that("step_select_vip, execution using top_p", {
  skip_if_not_installed("ranger")

  irisX <- iris[-5]
  y <- iris$Species

  base_model <- rand_forest(mode = "classification") %>%
    set_engine("ranger", importance = "permutation")

  rec <- iris %>%
    recipe(Species ~.) %>%
    step_select_vip(
      all_predictors(),
      outcome = "Species",
      model = base_model,
      top_p = 2
    )

  prepped <- prep(rec)
  selected <- juice(prepped)

  expect_length(names(selected), 3)
})


test_that("step_select_vip, execution using threshold", {
  skip_if_not_installed("ranger")

  irisX <- iris[-5]
  y <- iris$Species

  base_model <- rand_forest(mode = "classification") %>%
    set_engine("ranger", importance = "permutation")

  # test selection by retaining features with scores >= 50th percentile
  rec <- iris %>%
    recipe(Species ~.) %>%
    step_select_vip(
      all_predictors(),
      outcome = "Species",
      model = base_model,
      threshold = 0.5
    )

  prepped <- prep(rec)
  selected <- juice(prepped)

  expect_length(names(selected), 3)

  # test selection by retaining features with scores in 90th percentile
  rec <- iris %>%
    recipe(Species ~.) %>%
    step_select_vip(
      all_predictors(),
      outcome = "Species",
      model = base_model,
      threshold = 0.9
    )

  prepped <- prep(rec)
  selected <- juice(prepped)

  expect_length(names(selected), 2)
})