1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
|
"""
Testing for the gradient boosting module (sklearn.ensemble.gradient_boosting).
"""
import numpy as np
from numpy.testing import assert_array_equal
from numpy.testing import assert_equal
from nose.tools import assert_raises
from sklearn.metrics import mean_squared_error
from sklearn.utils import check_random_state
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.ensemble import GradientBoostingRegressor
from sklearn import datasets
# toy sample
X = [[-2, -1], [-1, -1], [-1, -2], [1, 1], [1, 2], [2, 1]]
y = [-1, -1, -1, 1, 1, 1]
T = [[-1, -1], [2, 2], [3, 2]]
true_result = [-1, 1, 1]
rng = np.random.RandomState(0)
# also load the boston dataset
# and randomly permute it
boston = datasets.load_boston()
perm = rng.permutation(boston.target.size)
boston.data = boston.data[perm]
boston.target = boston.target[perm]
# also load the iris dataset
# and randomly permute it
iris = datasets.load_iris()
perm = rng.permutation(iris.target.size)
iris.data = iris.data[perm]
iris.target = iris.target[perm]
def test_classification_toy():
"""Check classification on a toy dataset."""
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
assert_raises(ValueError, clf.predict, T)
clf.fit(X, y)
assert_array_equal(clf.predict(T), true_result)
assert_equal(100, len(clf.estimators_))
deviance_decrease = (clf.train_score_[:-1] - clf.train_score_[1:])
assert np.any(deviance_decrease >= 0.0), \
"Train deviance does not monotonically decrease."
def test_parameter_checks():
"""Check input parameter validation."""
assert_raises(ValueError, GradientBoostingClassifier, n_estimators=0)
assert_raises(ValueError, GradientBoostingClassifier, n_estimators=-1)
assert_raises(ValueError, GradientBoostingClassifier, learn_rate=0.0)
assert_raises(ValueError, GradientBoostingClassifier, learn_rate=-1.0)
assert_raises(ValueError, GradientBoostingRegressor, loss='foobar')
assert_raises(ValueError, GradientBoostingClassifier,
min_samples_split=0.0)
assert_raises(ValueError, GradientBoostingClassifier,
min_samples_split=-1.0)
assert_raises(ValueError, GradientBoostingClassifier, min_samples_leaf=0)
assert_raises(ValueError, GradientBoostingClassifier, min_samples_leaf=-1.)
assert_raises(ValueError, GradientBoostingClassifier, subsample=0.0)
assert_raises(ValueError, GradientBoostingClassifier, subsample=1.1)
assert_raises(ValueError, GradientBoostingClassifier, subsample=-0.1)
assert_raises(ValueError, GradientBoostingClassifier, max_depth=-0.1)
assert_raises(ValueError, GradientBoostingClassifier, max_depth=0)
assert_raises(ValueError, GradientBoostingClassifier, init={})
# test fit before feature importance
assert_raises(ValueError,
lambda: GradientBoostingClassifier().feature_importances_)
# binomial deviance requires ``n_classes == 2``.
assert_raises(ValueError,
lambda X, y: GradientBoostingClassifier(
loss='bdeviance').fit(X, y),
X, [0, 0, 1, 1, 2, 2])
# multinomial deviance requires ``n_classes > 2``.
assert_raises(ValueError,
lambda X, y: GradientBoostingClassifier(
loss='mdeviance').fit(X, y),
X, [0, 0, 1, 1, 1, 0])
# deviance requires ``n_classes >= 2``.
assert_raises(ValueError,
lambda X, y: GradientBoostingClassifier(
loss='deviance').fit(X, y),
X, [0, 0, 0, 0])
def test_classification_synthetic():
"""Test GradientBoostingClassifier on synthetic dataset used by
Hastie et al. in ESLII Example 12.7. """
X, y = datasets.make_hastie_10_2(n_samples=12000, random_state=1)
X_train, X_test = X[:2000], X[2000:]
y_train, y_test = y[:2000], y[2000:]
gbrt = GradientBoostingClassifier(n_estimators=100, min_samples_split=1,
max_depth=1,
learn_rate=1.0, random_state=0)
gbrt.fit(X_train, y_train)
error_rate = (1.0 - gbrt.score(X_test, y_test))
assert error_rate < 0.085, \
"GB failed with error %.4f" % error_rate
gbrt = GradientBoostingClassifier(n_estimators=200, min_samples_split=1,
max_depth=1,
learn_rate=1.0, subsample=0.5,
random_state=0)
gbrt.fit(X_train, y_train)
error_rate = (1.0 - gbrt.score(X_test, y_test))
assert error_rate < 0.08, \
"Stochastic GB failed with error %.4f" % error_rate
def test_boston():
"""Check consistency on dataset boston house prices with least squares
and least absolute deviation. """
for loss in ("ls", "lad"):
clf = GradientBoostingRegressor(n_estimators=100, loss=loss,
max_depth=4,
min_samples_split=1, random_state=1)
assert_raises(ValueError, clf.predict, boston.data)
clf.fit(boston.data, boston.target)
y_pred = clf.predict(boston.data)
mse = mean_squared_error(boston.target, y_pred)
assert mse < 6.0, "Failed with loss %s and mse = %.4f" % (loss, mse)
def test_iris():
"""Check consistency on dataset iris."""
for subsample in (1.0, 0.5):
clf = GradientBoostingClassifier(n_estimators=100, loss='deviance',
random_state=1, subsample=subsample)
clf.fit(iris.data, iris.target)
score = clf.score(iris.data, iris.target)
assert score > 0.9, "Failed with subsample %.1f " \
"and score = %f" % (subsample, score)
def test_regression_synthetic():
"""Test on synthetic regression datasets used in Leo Breiman,
`Bagging Predictors?. Machine Learning 24(2): 123-140 (1996). """
random_state = check_random_state(1)
regression_params = {'n_estimators': 100, 'max_depth': 4,
'min_samples_split': 1, 'learn_rate': 0.1,
'loss': 'ls'}
# Friedman1
X, y = datasets.make_friedman1(n_samples=1200,
random_state=random_state, noise=1.0)
X_train, y_train = X[:200], y[:200]
X_test, y_test = X[200:], y[200:]
clf = GradientBoostingRegressor()
clf.fit(X_train, y_train)
mse = mean_squared_error(y_test, clf.predict(X_test))
assert mse < 5.0, "Failed on Friedman1 with mse = %.4f" % mse
# Friedman2
X, y = datasets.make_friedman2(n_samples=1200, random_state=random_state)
X_train, y_train = X[:200], y[:200]
X_test, y_test = X[200:], y[200:]
clf = GradientBoostingRegressor(**regression_params)
clf.fit(X_train, y_train)
mse = mean_squared_error(y_test, clf.predict(X_test))
assert mse < 1700.0, "Failed on Friedman2 with mse = %.4f" % mse
# Friedman3
X, y = datasets.make_friedman3(n_samples=1200, random_state=random_state)
X_train, y_train = X[:200], y[:200]
X_test, y_test = X[200:], y[200:]
clf = GradientBoostingRegressor(**regression_params)
clf.fit(X_train, y_train)
mse = mean_squared_error(y_test, clf.predict(X_test))
assert mse < 0.015, "Failed on Friedman3 with mse = %.4f" % mse
def test_feature_importances():
clf = GradientBoostingRegressor(n_estimators=100, max_depth=4,
min_samples_split=1, random_state=1)
clf.fit(boston.data, boston.target)
feature_importances = clf.feature_importances_
# true feature importance ranking
true_ranking = np.array([3, 1, 8, 10, 2, 9, 4, 11, 0, 6, 7, 5, 12])
assert_array_equal(true_ranking, feature_importances.argsort())
def test_probability():
"""Predict probabilities."""
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
assert_raises(ValueError, clf.predict_proba, T)
clf.fit(X, y)
assert_array_equal(clf.predict(T), true_result)
# check if probabilities are in [0, 1].
y_proba = clf.predict_proba(T)
assert np.all(y_proba >= 0.0)
assert np.all(y_proba <= 1.0)
# derive predictions from probabilities
y_pred = clf.classes_.take(y_proba.argmax(axis=1), axis=0)
assert_array_equal(y_pred, true_result)
def test_check_inputs():
"""Test input checks (shape and type of X and y)."""
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
assert_raises(ValueError, clf.fit, X, y + [0, 1])
from scipy import sparse
X_sparse = sparse.csr_matrix(X)
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
assert_raises(ValueError, clf.fit, X_sparse, y)
clf = GradientBoostingClassifier().fit(X, y)
assert_raises(ValueError, clf.predict, X_sparse)
def test_check_inputs_predict():
"""X has wrong shape """
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
clf.fit(X, y)
x = np.array([1.0, 2.0])[:, np.newaxis]
assert_raises(ValueError, clf.predict, x)
x = np.array([])
assert_raises(ValueError, clf.predict, x)
x = np.array([1.0, 2.0, 3.0])[:, np.newaxis]
assert_raises(ValueError, clf.predict, x)
clf = GradientBoostingRegressor(n_estimators=100, random_state=1)
clf.fit(X, rng.rand(len(X)))
x = np.array([1.0, 2.0])[:, np.newaxis]
assert_raises(ValueError, clf.predict, x)
x = np.array([])
assert_raises(ValueError, clf.predict, x)
x = np.array([1.0, 2.0, 3.0])[:, np.newaxis]
assert_raises(ValueError, clf.predict, x)
def test_staged_predict():
"""Test whether staged decision function eventually gives
the same prediction.
"""
X, y = datasets.make_friedman1(n_samples=1200,
random_state=1, noise=1.0)
X_train, y_train = X[:200], y[:200]
X_test, y_test = X[200:], y[200:]
clf = GradientBoostingRegressor()
# test raise ValueError if not fitted
assert_raises(ValueError, lambda X: np.fromiter(
clf.staged_predict(X), dtype=np.float64), X_test)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
# test if prediction for last stage equals ``predict``
for y in clf.staged_predict(X_test):
assert_equal(y.shape, y_pred.shape)
assert_array_equal(y_pred, y)
def test_serialization():
"""Check model serialization."""
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
clf.fit(X, y)
assert_array_equal(clf.predict(T), true_result)
assert_equal(100, len(clf.estimators_))
try:
import cPickle as pickle
except ImportError:
import pickle
serialized_clf = pickle.dumps(clf, protocol=pickle.HIGHEST_PROTOCOL)
clf = None
clf = pickle.loads(serialized_clf)
assert_array_equal(clf.predict(T), true_result)
assert_equal(100, len(clf.estimators_))
def test_degenerate_targets():
"""Check if we can fit even though all targets are equal. """
clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
# classifier should raise exception
assert_raises(ValueError, clf.fit, X, np.ones(len(X)))
clf = GradientBoostingRegressor(n_estimators=100, random_state=1)
clf.fit(X, np.ones(len(X)))
clf.predict(rng.rand(2))
assert_array_equal(np.ones((1,), dtype=np.float64),
clf.predict(rng.rand(2)))
|
