Nothing
# Tests for caret's model trimming (R/trim.R: trim / trim.train).
# Shared train/test datasets (trim_class_*, trim_reg_*) live in helper-trim.R
test_that('glm classification', {
skip_on_cran()
set.seed(3162)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "glm",
tuneLength = 1,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(3162)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "glm",
tuneLength = 1,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('glm regression', {
skip_on_cran()
set.seed(7748)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "glm",
tuneLength = 1,
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(7748)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "glm",
tuneLength = 1,
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('glmnet classification', {
skip_on_cran()
skip_if_not_installed("glmnet")
set.seed(5290)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "glmnet",
tuneGrid = data.frame(lambda = 0.1, alpha = 0.5),
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(5290)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "glmnet",
tuneGrid = data.frame(lambda = 0.1, alpha = 0.5),
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('glmnet regression', {
skip_on_cran()
skip_if_not_installed("glmnet")
set.seed(6611)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "glmnet",
tuneGrid = data.frame(lambda = 0.1, alpha = 0.5),
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(6611)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "glmnet",
tuneGrid = data.frame(lambda = 0.1, alpha = 0.5),
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('rpart classification', {
skip_on_cran()
skip_if_not_installed("rpart")
set.seed(9521)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.22),
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(9521)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.22),
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('rpart regression', {
skip_on_cran()
skip_if_not_installed("rpart")
set.seed(1387)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.12),
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(1387)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.12),
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('rpart2 classification', {
skip_on_cran()
skip_if_not_installed("rpart")
set.seed(4476)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart2",
tuneGrid = data.frame(maxdepth = 3),
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(4476)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart2",
tuneGrid = data.frame(maxdepth = 3),
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('rpart2 regression', {
skip_on_cran()
skip_if_not_installed("rpart")
set.seed(8039)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart2",
tuneGrid = data.frame(maxdepth = 3),
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(8039)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart2",
tuneGrid = data.frame(maxdepth = 3),
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('treebag classification', {
skip_on_cran()
skip_if_not_installed("ipred")
set.seed(8124)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "treebag",
nbagg = 3,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(8124)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "treebag",
nbagg = 3,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('treebag regression', {
skip_on_cran()
skip_if_not_installed("ipred")
set.seed(2057)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "treebag",
nbagg = 3,
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(2057)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "treebag",
nbagg = 3,
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('bayesglm classification', {
skip_on_cran()
skip_if_not_installed("arm")
set.seed(4903)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "bayesglm",
tuneLength = 1,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(4903)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "bayesglm",
tuneLength = 1,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('bayesglm regression', {
skip_on_cran()
skip_if_not_installed("arm")
set.seed(1675)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "bayesglm",
tuneLength = 1,
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(1675)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "bayesglm",
tuneLength = 1,
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('single tree', {
skip_on_cran()
skip_if_not_installed("C50")
grid <- data.frame(
trials = 1,
model = "tree",
winnow = FALSE,
stringsAsFactors = TRUE
)
set.seed(7382)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(7382)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('single rule', {
skip_on_cran()
skip_if_not_installed("C50")
grid <- data.frame(
trials = 1,
model = "rules",
winnow = FALSE,
stringsAsFactors = TRUE
)
set.seed(2918)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(2918)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('boosted tree', {
skip_on_cran()
skip_if_not_installed("C50")
grid <- data.frame(
trials = 5,
model = "tree",
winnow = FALSE,
stringsAsFactors = TRUE
)
set.seed(5564)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(5564)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('boosted rule', {
skip_on_cran()
skip_if_not_installed("C50")
grid <- data.frame(
trials = 5,
model = "rules",
winnow = FALSE,
stringsAsFactors = TRUE
)
set.seed(9047)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
set.seed(9047)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "C5.0",
tuneGrid = grid,
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('train classification', {
skip_on_cran()
skip_if_not_installed("rpart")
skip_if_not_installed("ipred")
set.seed(3301)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.22),
preProc = c("center", "bagImpute"),
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
class_trim <- caret:::trim.train(class_trim)
set.seed(3301)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.22),
preProc = c("center", "bagImpute"),
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('train regression', {
skip_on_cran()
skip_if_not_installed("rpart")
set.seed(6720)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.12),
trControl = trainControl(method = "none", trim = TRUE)
)
reg_trim <- caret:::trim.train(reg_trim)
set.seed(6720)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "rpart",
tuneGrid = data.frame(cp = 0.12),
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
test_that('train/earth classification', {
skip_on_cran()
skip_if_not_installed("earth")
set.seed(1129)
class_trim <- train(
Class ~ .,
data = trim_class_tr,
method = "earth",
tuneGrid = data.frame(nprune = 3, degree = 1),
trControl = trainControl(method = "none", classProbs = TRUE, trim = TRUE)
)
class_trim <- caret:::trim.train(class_trim)
set.seed(1129)
class_notrim <- train(
Class ~ .,
data = trim_class_tr,
method = "earth",
tuneGrid = data.frame(nprune = 3, degree = 1),
trControl = trainControl(method = "none", classProbs = TRUE, trim = FALSE)
)
expect_equal(
predict(class_trim, trim_class_te),
predict(class_notrim, trim_class_te)
)
expect_equal(
predict(class_trim, trim_class_te, type = "prob"),
predict(class_notrim, trim_class_te, type = "prob")
)
expect_lt(object.size(class_trim), object.size(class_notrim))
})
test_that('train/earth regression', {
skip_on_cran()
skip_if_not_installed("earth")
set.seed(8456)
reg_trim <- train(
y ~ .,
data = trim_reg_tr,
method = "earth",
tuneGrid = data.frame(nprune = 3, degree = 1),
trControl = trainControl(method = "none", trim = TRUE)
)
set.seed(8456)
reg_notrim <- train(
y ~ .,
data = trim_reg_tr,
method = "earth",
tuneGrid = data.frame(nprune = 3, degree = 1),
trControl = trainControl(method = "none", trim = FALSE)
)
expect_equal(predict(reg_trim, trim_reg_te), predict(reg_notrim, trim_reg_te))
# NOTE: on current master the earth model has no `trim` method, so
# trainControl(trim = TRUE) does not shrink it. Re-enable this size check once
# the earth trim fix (separate PR, branch fix-earth-trim) is merged.
# expect_lt(object.size(reg_trim), object.size(reg_notrim))
})
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.