tests/testthat/test_LearnerRegr.R

In mlr3: Machine Learning in R - Next Generation

test_that("LearnerRegr predict_newdata_fast response works", {
  learner = lrn("regr.debug")
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$response)
})

test_that("LearnerRegr predict_newdata_fast se works", {
  learner = lrn("regr.debug", predict_type = "se")
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$se)
})

test_that("LearnerRegr predict_newdata_fast quantiles works", {
  learner = lrn("regr.debug", predict_type = "quantiles")
  learner$quantiles = c(0.1, 0.5, 0.9)
  learner$quantile_response = 0.5
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_matrix(pred$quantiles, nrows = nrow(newdata), ncols = length(learner$quantiles))
})

test_that("LearnerRegr predict_newdata_fast works with missing values", {
  learner = lrn("regr.debug", predict_missing = 0.5)
  learner$encapsulate("evaluate", fallback = lrn("regr.featureless"))
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$response, any.missing = FALSE)

  learner = lrn("regr.debug", predict_missing = 0.5, predict_type = "se")
  learner$encapsulate("evaluate", fallback = lrn("regr.featureless", predict_type = "se"))
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$se, any.missing = FALSE)

  learner = lrn("regr.debug", predict_missing = 0.5, predict_type = "quantiles")
  learner$quantiles = c(0.1, 0.5, 0.9)
  learner$quantile_response = 0.5
  fallback = lrn("regr.featureless", predict_type = "quantiles")
  fallback$quantiles = c(0.1, 0.5, 0.9)
  fallback$quantile_response = 0.5
  learner$encapsulate("evaluate", fallback = fallback)
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_matrix(pred$quantiles, nrows = nrow(newdata), ncols = length(learner$quantiles), any.missing = FALSE)
})

test_that("LearnerRegr predict_newdata_fast works with failed train", {
  learner = lrn("regr.debug", predict_missing = 0.5, error_train = 1)
  learner$encapsulate("evaluate", fallback = lrn("regr.featureless"))
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$response, any.missing = FALSE)

  learner = lrn("regr.debug", predict_missing = 0.5, predict_type = "se", error_train = 1)
  learner$encapsulate("evaluate", fallback = lrn("regr.featureless", predict_type = "se"))
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)

  expect_list(pred)
  expect_names(names(pred), subset.of = c("response", "se", "quantiles"))
  expect_numeric(pred$se, any.missing = FALSE)

  learner = lrn("regr.debug", predict_missing = 0.5, predict_type = "quantiles", error_train = 1)
  learner$quantiles = c(0.1, 0.5, 0.9)
  learner$quantile_response = 0.5
  fallback = lrn("regr.featureless", predict_type = "quantiles")
  fallback$quantiles = c(0.1, 0.5, 0.9)
  fallback$quantile_response = 0.5
  learner$encapsulate("evaluate", fallback = fallback)
  task = tsk("mtcars")
  newdata = task$data()
  learner$train(task)
  pred = learner$predict_newdata_fast(newdata)
})