tests/testthat/test-learn-params-awnb.R

In bnclassify: Learning Discrete Bayesian Network Classifiers from Data

context("AWNB") 

test_that("one tree", {
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(0)
  a <- awnb('class', car, bootstrap_size = 0.5, trees = 1)
  expect_equal(as.vector(a['buying']), 0.5773503, tolerance = 1e-5)
  expect_equal(as.vector(a['doors']), 0.3779645, tolerance = 1e-5)
  expect_equal(as.vector(a['persons']), 1)
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(0)
})

test_that("one two trees", {  
  a <- awnb('class', car, bootstrap_size = 0.5, trees = 2)
  expect_true(is_perm(names(a), colnames(car)[-7]))
  expect_equal(as.vector(a['buying']), 0.5773503, tolerance = 1e-5)
  expect_equal(as.vector(a['doors']), 0.3931064, tolerance = 1e-5)
  expect_equal(as.vector(a['persons']), 1) 
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(0)
  a <- awnb('Class', voting, bootstrap_size = 0.5, trees = 10)
  expect_true(is_perm(names(a), colnames(voting)[-17]))
  expect_equal(as.vector(a['superfund_right_to_sue']), 0.21019141, tolerance = 1e-5)
  expect_equal(as.vector(a['mx_missile']), 0.30487476, tolerance = 1e-5)
  expect_equal(as.vector(a['immigration']), 0.39963413)
})

test_that("one no tree", {
  expect_error(awnb('Class', dataset = voting[1:2, ],  
                    bootstrap_size = 0.5, trees = 10), "empty")
})

test_that("weights for features not in tree", {
  suppressWarnings(RNGversion("3.5.0"))
  set.seed(0)
  a <- awnb('class', car[sample(1:1000, 10), , drop = FALSE], 
            bootstrap_size = 1, trees = 1)
  a
  expect_true(is_perm(names(a), colnames(car)[-7]))
})