tests/testthat/test-ngb_reg_class.R

In Akai01/ngboost: R Interface to Ngboost

test_that("NGBRegression", {

  if_not_ngboost_exist_skip()

  library(ngboost)
  set.seed(2022)

  y_train = rnorm(100, 10, 1)

  x_train <- data.frame(x1 = rnorm(100, 15, 1),
                        x2 = rnorm(100, 14, 1),
                        x3 = rnorm(100, 13, 1))

  x_val <- data.frame(x1 = rnorm(20, 15, 1),
                      x2 = rnorm(20, 14, 1),
                      x3 = rnorm(20, 13, 1))
  y_val = rnorm(20, 10, 1)

  model <- NGBRegression$new(Dist = Dist(),
                             Base= sklearner(),
                             Score = Scores(),
                             natural_gradient=TRUE,
                             n_estimators= 100,
                             learning_rate= 0.002,
                             minibatch_frac= 0.8,
                             col_sample= 0.9,
                             verbose=FALSE,
                             verbose_eval=100,
                             tol=1e-5, random_state = 42L)

  model$fit(X = x_train, Y = y_train, X_val = x_val, Y_val = y_val)

  the_dist <- model$pred_dist(x_val)

  expect_equal(model$predict(x_val)[1], 10.05477, tolerance = 0.5)
})

test_that("NGBClassifier", {

  if_not_ngboost_exist_skip()

  library(ngboost)

  set.seed(2022)
  y_train = rpois(100, 1)

  x_train <- data.frame(x1 = rnorm(100, 15, 1),
                        x2 = rnorm(100, 14, 1),
                        x3 = rnorm(100, 13, 1))

  x_val <- data.frame(x1 = rnorm(20, 15, 1),
                      x2 = rnorm(20, 14, 1),
                      x3 = rnorm(20, 13, 1))
  y_val = rpois(20, 1)


  model <- NGBClassifier$new(Dist = Dist("k_categorical", k = 6),
                             Base = sklearner(),
                             Score = Scores("LogScore"),
                             natural_gradient = TRUE,
                             n_estimators = 100,
                             tol = 1e-5,
                             random_state = NULL)

  model$fit(X = x_train, Y = y_train, X_val = x_val, Y_val = y_val)

  model$predict(x_val)[1]

  the_dist <- model$pred_dist(x_val)

  expect_equal(model$predict(x_val)[1], 0,)
})