test_that("get same prediction twice via set.seed for pseudo-rng seeding", { set.seed(42) ind <-1:150 %in% sample(150, 100) set.seed(2) rf1 <- train(data=iris[ind, ], response_name="Species") pred1 <- predict(rf1, newdata=iris[!ind, ]) set.seed(2) rf2 <- train(data=iris[ind, ], response_name="Species") pred2 <- predict(rf2, newdata=iris[!ind, ]) expect_equal(pred1$predictions, pred2$predictions) }) test_that("get same prediction twice via 'seed' argument for pseudo-rng seeding", { set.seed(42) ind <-1:150 %in% sample(150, 100) rf1 <- train(data=iris[ind, ], response_name="Species", seed=1) pred1 <- predict(rf1, newdata=iris[!ind, ], seed=2) rf2 <- train(data=iris[ind, ], response_name="Species", seed=1) pred2 <- predict(rf2, newdata=iris[!ind, ], seed=2) expect_equal(pred1$predictions, pred2$predictions) })