def test_learner_scorer_previous_transformation(self):
learner = LogisticRegressionLearner()
from Orange.preprocess import Discretize
data = Discretize()(self.iris)
scores = learner.score_data(data)
# scores should be defined and positive
self.assertTrue(np.all(scores > 0))
def test_learner_scorer(self):
data = Table('voting')
learner = LogisticRegressionLearner()
scores = learner.score_data(data)
self.assertEqual('physician-fee-freeze',
data.domain.attributes[np.argmax(scores)].name)
self.assertEqual(len(scores), len(data.domain.attributes))
def test_learner_scorer_multiclass(self):
attr = self.zoo.domain.attributes
learner = LogisticRegressionLearner()
scores = learner.score_data(self.zoo)
self.assertEqual('aquatic', attr[np.argmax(scores[0])].name) # amphibian
self.assertEqual('feathers', attr[np.argmax(scores[1])].name) # bird
self.assertEqual('fins', attr[np.argmax(scores[2])].name) # fish
self.assertEqual('legs', attr[np.argmax(scores[3])].name) # insect
self.assertEqual('backbone', attr[np.argmax(scores[4])].name) # invertebrate
self.assertEqual('milk', attr[np.argmax(scores[5])].name) # mammal
self.assertEqual('hair', attr[np.argmax(scores[6])].name) # reptile
self.assertEqual(scores.shape,
(len(self.zoo.domain.class_var.values), len(attr)))
def test_learner_scorer_multiclass(self):
attr = self.zoo.domain.attributes
learner = LogisticRegressionLearner()
scores = learner.score_data(self.zoo)
self.assertEqual('aquatic', attr[np.argmax(scores[0])].name)
self.assertEqual('feathers', attr[np.argmax(scores[1])].name)
self.assertEqual('fins', attr[np.argmax(scores[2])].name)
self.assertEqual('backbone', attr[np.argmax(scores[3])].name)
self.assertEqual('backbone', attr[np.argmax(scores[4])].name)
self.assertEqual('milk', attr[np.argmax(scores[5])].name)
self.assertEqual('hair', attr[np.argmax(scores[6])].name)
self.assertEqual(scores.shape,
(len(self.zoo.domain.class_var.values), len(attr)))
def test_bad_learner(self):
"""
Some learners on input might raise error.
GH-38
"""
w = self.widget
self.assertFalse(w.Error.fitting_failed.is_shown())
learner = LogisticRegressionLearner()
learner.preprocessors = [Discretize()]
self.send_signal(w.Inputs.learner, learner)
self.send_signal(w.Inputs.data, self.iris)
self.assertTrue(w.Error.fitting_failed.is_shown())
learner.preprocessors = []
self.send_signal(w.Inputs.learner, learner)
self.assertFalse(w.Error.fitting_failed.is_shown())
def test_learner_scorer(self):
learner = LogisticRegressionLearner()
scores = learner.score_data(self.heart_disease)
self.assertEqual('major vessels colored',
self.heart_disease.domain.attributes[np.argmax(scores)].name)
self.assertEqual(scores.shape, (1, len(self.heart_disease.domain.attributes)))
def test_LogisticRegressionNormalization_todo(self):
with self.assertRaises(TypeError):
lr = LogisticRegressionLearner(normalize=True)
def test_predict_on_instance(self):
lr = LogisticRegressionLearner()
m = lr(self.zoo)
probs = m(self.zoo[50], m.Probs)
probs2 = m(self.zoo[50, :], m.Probs)
np.testing.assert_almost_equal(probs, probs2[0])
def test_learner_scorer_multiclass_feature(self):
learner = LogisticRegressionLearner()
scores = learner.score_data(self.zoo)
for i, attr in enumerate(self.zoo.domain.attributes):
score = learner.score_data(self.zoo, attr)
np.testing.assert_array_almost_equal(score, scores[:, i])
def test_nomogram_lr_multiclass(self):
"""Check probabilities for logistic regression classifier for various
values of classes and radio buttons for multiclass data"""
cls = LogisticRegressionLearner()(self.lenses)
self._test_helper(cls, [9, 45, 52])
def test_predict_savgov_same_domain(self):
data = SavitzkyGolayFiltering(window=9, polyorder=2,
deriv=2)(self.collagen)
train, test = separate_learn_test(data)
auc = AUC(TestOnTestData()(train, test, [LogisticRegressionLearner()]))
self.assertGreater(auc, 0.85)
def test_coefficients(self):
data = Table("voting")
learn = LogisticRegressionLearner()
model = learn(data)
coef = model.coefficients
self.assertEqual(len(coef[0]), len(model.domain.attributes))
def test_probability(self):
table = Table('iris')
learn = LogisticRegressionLearner(penalty='l1')
clf = learn(table[:100])
p = clf(table[100:], ret=Model.Probs)
self.assertTrue(all(abs(p.sum(axis=1) - 1) < 1e-6))
""" Documentation script """
from Orange.classification import LogisticRegressionLearner
from Orange.evaluation.testing import CrossValidation
from Orange.evaluation.scoring import AUC
from orangecontrib.bioinformatics.geo.dataset import GDS
gds = GDS("GDS2960")
data = gds.get_data(sample_type="disease state", transpose=True, report_genes=True)
print("Samples: %d, Genes: %d" % (len(data), len(data.domain.attributes)))
learners = [LogisticRegressionLearner()]
results = CrossValidation(data, learners, k=10)
print("AUC = %.3f" % AUC(results)[0])
def test_logreg(self):
self.assertTrue(LogisticRegressionLearner().supports_weights,
"Either LogisticRegression no longer supports weighted tables"
"or SklLearner.supports_weights is out-of-date.")
def test_multinomial(self):
table = Table("titanic")
lr = LogisticRegressionLearner()
assert isinstance(lr, Orange.classification.SklLearner)
res = CrossValidation(table, [lr], k=2)
self.assertTrue(0.7 < Orange.evaluation.AUC(res)[0] < 0.9)
def test_LogisticRegression(self):
table = Orange.data.Table('iris')
learn = LogisticRegressionLearner()
results = Orange.evaluation.CrossValidation(table, [learn], k=2)
ca = Orange.evaluation.CA(results)
self.assertTrue(0.9 < ca < 1.0)
from Orange.classification import LogisticRegressionLearner
from Orange.data import ContinuousVariable
from Orange.evaluation.testing import TestOnTestData
from Orange.evaluation.scoring import AUC
from orangecontrib.spectroscopy.tests.test_preprocess import \
PREPROCESSORS_INDEPENDENT_SAMPLES, \
PREPROCESSORS
from orangecontrib.spectroscopy.tests.test_preprocess import SMALL_COLLAGEN, preprocessor_data
from orangecontrib.spectroscopy.preprocess import Interpolate, \
Cut, SavitzkyGolayFiltering
from orangecontrib.spectroscopy.data import getx
logreg = LogisticRegressionLearner(max_iter=1000)
def separate_learn_test(data):
sf = ms.ShuffleSplit(n_splits=1,
test_size=0.2,
random_state=np.random.RandomState(0))
(traini, testi), = sf.split(y=data.Y, X=data.X)
return data[traini], data[testi]
def slightly_change_wavenumbers(data, change):
natts = [
ContinuousVariable(float(a.name) + change)
for a in data.domain.attributes
]
def __init__(self, learners, aggregate=LogisticRegressionLearner(), k=5):
super().__init__(learners=learners, aggregate=aggregate, k=k)
def test_LogisticRegression(self):
table = Table('iris')
learn = LogisticRegressionLearner()
results = CrossValidation(table, [learn], k=2)
ca = CA(results)
self.assertTrue(0.8 < ca < 1.0)
def test_predict_same_domain(self):
train, test = separate_learn_test(self.collagen)
auc = AUC(TestOnTestData()(train, test, [LogisticRegressionLearner()]))
self.assertGreater(auc, 0.9) # easy dataset
def test_LogisticRegression(self):
learn = LogisticRegressionLearner()
results = CrossValidation(self.voting, [learn], k=2)
ca = CA(results)
self.assertGreater(ca, 0.8)
self.assertLess(ca, 1.0)
def setUp(self):
self.widget = self.create_widget(OWPythonScript)
self.iris = Table("iris")
self.learner = LogisticRegressionLearner()
self.model = self.learner(self.iris)
def test_learner_scorer(self):
learner = LogisticRegressionLearner()
scores = learner.score_data(self.voting)
self.assertEqual('physician-fee-freeze',
self.voting.domain.attributes[np.argmax(scores)].name)
self.assertEqual(scores.shape, (1, len(self.voting.domain.attributes)))
def test_coefficients(self):
learn = LogisticRegressionLearner()
model = learn(self.heart_disease)
coef = model.coefficients
self.assertEqual(len(coef[0]), len(model.domain.attributes))
from Orange.classification import LogisticRegressionLearner, NaiveBayesLearner, \
RandomForestLearner
import orangecontrib.evcrules.rules as rules
from orangecontrib.evcrules.logistic import LRRulesLearner
datasets = [
'ionosphere', 'adult_sample', 'iris', 'breast-cancer', 'bupa', 'titanic'
]
for d in datasets:
data = Table(d)
rule_learner = rules.RulesStar(evc=True,
add_sub_rules=True,
parent_alpha=0.5)
rule_learner_m = rules.RulesStar(evc=False,
m=22,
add_sub_rules=True,
parent_alpha=0.5)
# compare lr with rules, lr without rules and sklearn's lr
learners = [
LRRulesLearner(opt_penalty=True, rule_learner=rule_learner),
LRRulesLearner(opt_penalty=True, rule_learner=rule_learner_m),
LRRulesLearner(opt_penalty=True),
LogisticRegressionLearner(C=1),
NaiveBayesLearner(),
RandomForestLearner()
]
res = CrossValidation(data, learners, k=5)
print("Dataset: ", d)
for l, ca, auc, ll in zip(learners, CA(res), AUC(res), LogLoss(res)):
print("learner: {}\nCA: {}\nAUC: {}\n LL: {}".format(l, ca, auc, ll))
def test_single_class(self):
t = self.iris[60:90]
self.assertEqual(len(np.unique(t.Y)), 1)
learn = LogisticRegressionLearner()
with self.assertRaises(ValueError):
learn(t)
def test_multiclass_auc_multi_learners(self):
learners = [LogisticRegressionLearner(), MajorityLearner()]
res = CrossValidation(self.iris, learners, k=10)
self.assertGreater(AUC(res)[0], 0.6)
self.assertLess(AUC(res)[1], 0.6)
self.assertGreater(AUC(res)[1], 0.4)
def test_probability(self):
learn = LogisticRegressionLearner(penalty='l1')
clf = learn(self.iris[:100])
p = clf(self.iris[100:], ret=Model.Probs)
self.assertLess(abs(p.sum(axis=1) - 1).all(), 1e-6)
def test_precision(self):
learner = LogisticRegressionLearner(preprocessors=[])
results = TestOnTrainingData(self.data, [learner])
self.assertAlmostEqual(Precision(results)[0], 0.962, 3)
def test_predict_samename_domain(self):
train, test = separate_learn_test(self.collagen)
test = destroy_atts_conversion(test)
aucdestroyed = AUC(
TestOnTestData(train, test, [LogisticRegressionLearner()]))
self.assertTrue(0.45 < aucdestroyed < 0.55)
def test_learner_scorer_multiclass_feature(self):
learner = LogisticRegressionLearner()
scores = learner.score_data(self.zoo)
for i, attr in enumerate(self.zoo.domain.attributes):
score = learner.score_data(self.zoo, attr)
np.testing.assert_array_almost_equal(score, scores[:, i])
def test_learner_scorer(self):
learner = LogisticRegressionLearner()
scores = learner.score_data(self.voting)
self.assertEqual('physician-fee-freeze',
self.voting.domain.attributes[np.argmax(scores)].name)
self.assertEqual(scores.shape, (1, len(self.voting.domain.attributes)))
def test_logreg(self):
self.assertFalse(
LogisticRegressionLearner().supports_weights,
"Logistic regression has its supports_weights overridden because "
"liblinear doesn't support them (even though the parameter exists)"
)
def test_predict_different_domain(self):
train, test = separate_learn_test(self.collagen)
test = Interpolate(points=getx(test) - 1)(test) # other test domain
aucdestroyed = AUC(
TestOnTestData(train, test, [LogisticRegressionLearner()]))
self.assertTrue(0.45 < aucdestroyed < 0.55)
def __init__(self,
learners,
aggregate=LogisticRegressionLearner(),
k=5,
preprocessors=None):
super().__init__(learners, aggregate, k=k, preprocessors=preprocessors)
def test_unknown(self):
table = Table("iris")
tree = LogisticRegressionLearner()(table)
tree([1, 2, None])
[(data_attributes[i].name, dat.x[i]) for i in positive_idx],
[(data_attributes[i].name, dat.x[i]) for i in negative_ids],
)
)
pos_segments = _compute_segments(sv[positive_idx], pred[target_class])
neg_segments = _compute_segments(sv[negative_ids], pred[target_class])
segments.append((pos_segments, neg_segments))
ranges.append(
(
pos_segments[-1][1]
if len(pos_segments)
else pred[target_class],
neg_segments[-1][1]
if len(neg_segments)
else pred[target_class],
)
)
return selected_shap_values, segments, selected_labels, ranges
if __name__ == "__main__":
from Orange.classification import LogisticRegressionLearner
data_ = Table.from_file("heart_disease.tab")
learner = LogisticRegressionLearner()
model_ = learner(data_)
shap_val, transformed_domain, mask, colors_ = get_shap_values_and_colors(
model_, data_
)