def test_importances():
"""Check variable importances."""
X, y = datasets.make_classification(n_samples=2000,
n_features=10,
n_informative=3,
n_redundant=0,
n_repeated=0,
shuffle=False,
random_state=0)
for name, Tree in CLF_TREES.items():
clf = Tree(random_state=0)
clf.fit(X, y)
importances = clf.feature_importances_
n_important = np.sum(importances > 0.1)
assert_equal(importances.shape[0], 10, "Failed with {0}".format(name))
assert_equal(n_important, 3, "Failed with {0}".format(name))
X_new = clf.transform(X, threshold="mean")
assert_less(0, X_new.shape[1], "Failed with {0}".format(name))
assert_less(X_new.shape[1], X.shape[1], "Failed with {0}".format(name))
# Check on iris that importances are the same for all builders
clf = DecisionTreeClassifier(random_state=0)
clf.fit(iris.data, iris.target)
clf2 = DecisionTreeClassifier(random_state=0,
max_leaf_nodes=len(iris.data))
clf2.fit(iris.data, iris.target)
assert_array_equal(clf.feature_importances_,
clf2.feature_importances_)
def test_importances():
"""Check variable importances."""
X, y = datasets.make_classification(n_samples=2000,
n_features=10,
n_informative=3,
n_redundant=0,
n_repeated=0,
shuffle=False,
random_state=0)
for name, Tree in CLF_TREES.items():
clf = Tree(random_state=0)
clf.fit(X, y)
importances = clf.feature_importances_
n_important = np.sum(importances > 0.1)
assert_equal(importances.shape[0], 10, "Failed with {0}".format(name))
assert_equal(n_important, 3, "Failed with {0}".format(name))
X_new = clf.transform(X, threshold="mean")
assert_less(0, X_new.shape[1], "Failed with {0}".format(name))
assert_less(X_new.shape[1], X.shape[1], "Failed with {0}".format(name))
# Check on iris that importances are the same for all builders
clf = DecisionTreeClassifier(random_state=0)
clf.fit(iris.data, iris.target)
clf2 = DecisionTreeClassifier(random_state=0,
max_leaf_nodes=len(iris.data))
clf2.fit(iris.data, iris.target)
assert_array_equal(clf.feature_importances_, clf2.feature_importances_)
def test_big_input():
"""Test if the warning for too large inputs is appropriate."""
X = np.repeat(10 ** 40., 4).astype(np.float64).reshape(-1, 1)
clf = DecisionTreeClassifier()
try:
clf.fit(X, [0, 1, 0, 1])
except ValueError as e:
assert_in("float32", str(e))
def test_big_input():
"""Test if the warning for too large inputs is appropriate."""
X = np.repeat(10**40., 4).astype(np.float64).reshape(-1, 1)
clf = DecisionTreeClassifier()
try:
clf.fit(X, [0, 1, 0, 1])
except ValueError as e:
assert_in("float32", str(e))
def test_importances_variance_equal_mse():
"""Check that gini is equivalent to mse for binary output variable"""
from sklearn.tree._tree import TREE_LEAF
X, y = datasets.make_classification(n_samples=2000,
n_features=10,
n_informative=3,
n_redundant=0,
n_repeated=0,
shuffle=False,
random_state=0)
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1)
var = DecisionTreeClassifier(criterion="variance",
random_state=0).fit(X_train, y_train)
gini = DecisionTreeClassifier(criterion="gini",
random_state=0).fit(X_train, y_train)
reg = DecisionTreeRegressor(criterion="mse",
random_state=0).fit(X_train, y_train)
gini_leaves = gini.tree_.children_left == TREE_LEAF
var_leaves = var.tree_.children_left == TREE_LEAF
assert_array_equal(var.tree_.feature, reg.tree_.feature)
assert_almost_equal(var.feature_importances_, reg.feature_importances_)
assert_array_equal(var.tree_.children_left, reg.tree_.children_left)
assert_array_equal(var.tree_.children_right, reg.tree_.children_right)
assert_array_equal(var.tree_.n_node_samples, reg.tree_.n_node_samples)
assert_array_equal(var.tree_.feature, gini.tree_.feature)
assert_almost_equal(var.feature_importances_, gini.feature_importances_)
assert_array_equal(var.tree_.children_left, gini.tree_.children_left)
assert_array_equal(var.tree_.children_right, gini.tree_.children_right)
assert_array_equal(var.tree_.n_node_samples, gini.tree_.n_node_samples)
assert_almost_equal(var.tree_.value[var_leaves], gini.tree_.value[gini_leaves])
clf = DecisionTreeClassifier(criterion="gini", random_state=0,
output_transformer=IdentityProjection(),
).fit(X_train, y_train)
clf_leaves = clf.tree_.children_left == TREE_LEAF
assert_array_equal(clf.tree_.feature, reg.tree_.feature)
assert_almost_equal(clf.feature_importances_, reg.feature_importances_)
assert_array_equal(clf.tree_.children_left, reg.tree_.children_left)
assert_array_equal(clf.tree_.children_right, reg.tree_.children_right)
assert_array_equal(clf.tree_.n_node_samples, reg.tree_.n_node_samples)
assert_array_equal(clf.tree_.n_node_samples, reg.tree_.n_node_samples)
assert_array_equal(clf.tree_.feature, gini.tree_.feature)
assert_almost_equal(clf.feature_importances_, gini.feature_importances_)
assert_array_equal(clf.tree_.children_left, gini.tree_.children_left)
assert_array_equal(clf.tree_.children_right, gini.tree_.children_right)
assert_array_equal(clf.tree_.n_node_samples, gini.tree_.n_node_samples)
assert_almost_equal(clf.tree_.value[clf_leaves], gini.tree_.value[gini_leaves])
def test_arrays_persist():
"""Ensure property arrays' memory stays alive when tree disappears
non-regression for #2726
"""
for attr in [
'n_classes', 'value', 'children_left', 'children_right',
'threshold', 'impurity', 'feature', 'n_node_samples'
]:
value = getattr(DecisionTreeClassifier().fit([[0]], [0]).tree_, attr)
# if pointing to freed memory, contents may be arbitrary
assert_true(-2 <= value.flat[0] < 2,
'Array points to arbitrary memory')
def test_sample_weight_invalid():
"""Check sample weighting raises errors."""
X = np.arange(100)[:, np.newaxis]
y = np.ones(100)
y[:50] = 0.0
clf = DecisionTreeClassifier(random_state=0)
sample_weight = np.random.rand(100, 1)
assert_raises(ValueError, clf.fit, X, y, sample_weight=sample_weight)
sample_weight = np.array(0)
assert_raises(ValueError, clf.fit, X, y, sample_weight=sample_weight)
sample_weight = np.ones(101)
assert_raises(ValueError, clf.fit, X, y, sample_weight=sample_weight)
sample_weight = np.ones(99)
assert_raises(ValueError, clf.fit, X, y, sample_weight=sample_weight)
def test_importances_gini_equal_mse():
"""Check that gini is equivalent to mse for binary output variable"""
X, y = datasets.make_classification(n_samples=2000,
n_features=10,
n_informative=3,
n_redundant=0,
n_repeated=0,
shuffle=False,
random_state=0)
# The gini index and the mean square error (variance) might differ due
# to numerical instability. Since those instabilities mainly occurs at
# high tree depth, we restrict this maximal depth.
clf = DecisionTreeClassifier(criterion="gini", max_depth=5,
random_state=0).fit(X, y)
reg = DecisionTreeRegressor(criterion="mse", max_depth=5,
random_state=0).fit(X, y)
assert_almost_equal(clf.feature_importances_, reg.feature_importances_)
assert_array_equal(clf.tree_.feature, reg.tree_.feature)
assert_array_equal(clf.tree_.children_left, reg.tree_.children_left)
assert_array_equal(clf.tree_.children_right, reg.tree_.children_right)
assert_array_equal(clf.tree_.n_node_samples, reg.tree_.n_node_samples)
def test_sample_weight():
"""Check sample weighting."""
# Test that zero-weighted samples are not taken into account
X = np.arange(100)[:, np.newaxis]
y = np.ones(100)
y[:50] = 0.0
sample_weight = np.ones(100)
sample_weight[y == 0] = 0.0
clf = DecisionTreeClassifier(random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_array_equal(clf.predict(X), np.ones(100))
# Test that low weighted samples are not taken into account at low depth
X = np.arange(200)[:, np.newaxis]
y = np.zeros(200)
y[50:100] = 1
y[100:200] = 2
X[100:200, 0] = 200
sample_weight = np.ones(200)
sample_weight[y == 2] = .51 # Samples of class '2' are still weightier
clf = DecisionTreeClassifier(max_depth=1, random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_equal(clf.tree_.threshold[0], 149.5)
sample_weight[y == 2] = .5 # Samples of class '2' are no longer weightier
clf = DecisionTreeClassifier(max_depth=1, random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_equal(clf.tree_.threshold[0], 49.5) # Threshold should have moved
# Test that sample weighting is the same as having duplicates
X = iris.data
y = iris.target
duplicates = rng.randint(0, X.shape[0], 200)
clf = DecisionTreeClassifier(random_state=1)
clf.fit(X[duplicates], y[duplicates])
sample_weight = np.bincount(duplicates, minlength=X.shape[0])
clf2 = DecisionTreeClassifier(random_state=1)
clf2.fit(X, y, sample_weight=sample_weight)
internal = clf.tree_.children_left != tree._tree.TREE_LEAF
assert_array_almost_equal(clf.tree_.threshold[internal],
clf2.tree_.threshold[internal])
def test_sample_weight():
"""Check sample weighting."""
# Test that zero-weighted samples are not taken into account
X = np.arange(100)[:, np.newaxis]
y = np.ones(100)
y[:50] = 0.0
sample_weight = np.ones(100)
sample_weight[y == 0] = 0.0
clf = DecisionTreeClassifier(random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_array_equal(clf.predict(X), np.ones(100))
# Test that low weighted samples are not taken into account at low depth
X = np.arange(200)[:, np.newaxis]
y = np.zeros(200)
y[50:100] = 1
y[100:200] = 2
X[100:200, 0] = 200
sample_weight = np.ones(200)
sample_weight[y == 2] = .51 # Samples of class '2' are still weightier
clf = DecisionTreeClassifier(max_depth=1, random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_equal(clf.tree_.threshold[0], 149.5)
sample_weight[y == 2] = .5 # Samples of class '2' are no longer weightier
clf = DecisionTreeClassifier(max_depth=1, random_state=0)
clf.fit(X, y, sample_weight=sample_weight)
assert_equal(clf.tree_.threshold[0], 49.5) # Threshold should have moved
# Test that sample weighting is the same as having duplicates
X = iris.data
y = iris.target
duplicates = rng.randint(0, X.shape[0], 200)
clf = DecisionTreeClassifier(random_state=1)
clf.fit(X[duplicates], y[duplicates])
sample_weight = np.bincount(duplicates, minlength=X.shape[0])
clf2 = DecisionTreeClassifier(random_state=1)
clf2.fit(X, y, sample_weight=sample_weight)
internal = clf.tree_.children_left != tree._tree.TREE_LEAF
assert_array_almost_equal(clf.tree_.threshold[internal],
clf2.tree_.threshold[internal])
def test_importances_raises():
"""Check if variable importance before fit raises ValueError. """
clf = DecisionTreeClassifier()
clf.feature_importances_