def set_learner(self):
self.learner = tree.TreeLearner(
criterion=self.scores[self.attribute_score][1],
max_depth=self.max_depth,
min_samples_split=self.min_internal,
min_samples_leaf=self.min_leaf,
preprocessors=self.preprocessors)
self.learner.name = self.model_name
self.send("Learner", self.learner)
self.error(1)
if self.data is not None:
try:
self.classifier = self.learner(self.data)
self.classifier.name = self.model_name
self.classifier.instances = self.data
except Exception as errValue:
self.error(1, str(errValue))
self.classifier = None
else:
self.classifier = None
self.send("Classification Tree", self.classifier)
class TestClassification(testing.LearnerTestCase):
LEARNER = ctree.TreeLearner(max_depth=50)
def setUp(self):
from Orange.ensemble import stacking
from Orange.classification import tree, bayes
self.learner = stacking.StackedClassificationLearner(
[tree.TreeLearner(), bayes.NaiveLearner()], folds=5)
# NOTE: Needed by the save_treegraph_image() function.
import sys
from OWWidget import QApplication
a = QApplication(sys.argv)
# find out the current file's location so it can be used to compute the
# location of other files/directories
cur_dir = os.path.dirname(os.path.abspath(__file__))
path_prefix = os.path.abspath(os.path.join(cur_dir, "../../"))
base_learners = OrderedDict()
import Orange.classification.tree as octree
import Orange.feature.scoring as fscoring
split_const = octree.SplitConstructor_ExhaustiveBinary(
measure=fscoring.InfoGain())
base_learners["orange_tree"] = octree.TreeLearner(split=split_const,
store_instances=True)
base_learners["orange_tree_pruned"] = octree.TreeLearner(
split=split_const,
min_instances=10,
same_majority_pruning=True,
store_instances=True)
measures = []
measures.append("CA")
measures.append("AUC")
learners = OrderedDict()
learners["Tree"] = bin_exp.TreeMarkedAndMergedLearner()
learners["ForcedTree"] = bin_exp.ForcedFirstSplitMTLLearner(
first_split_attr="id")
no_filter = prefiltering.NoFilter()
for subset in subsets:
if subset != None:
subtree = bl(subset, weight)
root_node.branches.append(subtree.tree)
# create an (internal) Orange.core.TreeClassifier object
descender = getattr(self, "descender",
octree.Descender_UnknownMergeAsBranchSizes())
tree = octree._TreeClassifier(domain=instances.domain,
tree=root_node,
descender=descender)
# perform post pruning
if getattr(self, "same_majority_pruning", 0):
tree = Pruner_SameMajority(tree)
if getattr(self, "m_pruning", 0):
tree = Pruner_m(tree, m=self.m_pruning)
return octree.TreeClassifier(base_classifier=tree)
if __name__ == "__main__":
data = Orange.data.Table("titanic")
nt = octree.TreeLearner(data)
print "'Normal' TreeLearner:"
print nt # should have the 'sex' attribute as the first split
print
ffst = ForcedFirstSplitTreeLearner(data, first_split_attr="age")
print "ForcedFirstSplitTreeLearner:"
print ffst # should have 'age' attribute as the first split
"input. Model n_features is {} and input "
"n_features is {}".format(self.n_features_,
n_features))
# convert numpy data to Orange
orange_test_data = convert_numpy_data_to_orange(self.orange_domain_, X)
# classify all examples with the previously built classifier
p = np.empty((n_samples, self.n_classes_))
for i, ex in enumerate(orange_test_data):
p[i, :] = list(
self.orange_classifier_(ex, Orange.core.GetProbabilities))
return p
if __name__ == "__main__":
data = Orange.data.Table("titanic")
nt = octree.TreeLearner(data)
print "'Normal' TreeLearner:"
print nt # should have the 'sex' attribute as the first split
print
ffst = ForcedFirstSplitTreeLearner(data, first_split_attr="age")
print "ForcedFirstSplitTreeLearner (with the first split forced to 'age'):"
print ffst # should have 'age' attribute as the first split
print
import PyMTL.synthetic_data as sd
a, d = 8, 4
attr, func = sd.generate_boolean_function(a, d, random_seed=2)
print "Boolean function (a={}, d={}): {}".format(a, d, func)
X, y = sd.generate_examples(attr, func, n=100, random_state=10)
print "% of True values in y: {:.2f}".format(100 * sum(y == True) / len(y))
if __name__ == "__main__":
import os.path
# find out the current file's location so it can be used to compute the
# location of other files/directories
cur_dir = os.path.dirname(os.path.abspath(__file__))
path_prefix = os.path.abspath(os.path.join(cur_dir, "../../"))
results_path = os.path.join(path_prefix,
"results/binarization_experiment/")
import Orange.classification.tree as octree
import Orange.feature.scoring as fscoring
split_const = octree.SplitConstructor_ExhaustiveBinary(
measure=fscoring.InfoGain())
tree_learner = octree.TreeLearner(split=split_const,
min_instances=10,
same_majority_pruning=True,
store_instances=True)
from Orange.data import Table
# TEST for equality of "original" vs. "pickled/unpickled" Orange trees
from PyMTL.util import pickle_obj, unpickle_obj
import numpy as np
for i in range(10):
data = Table(
os.path.join(
results_path, "bool_func-a8d4n100g2tg5nse0.0rs15"
"nls10-seed63-complete_test/orange_merged_learn-"
"repetition{}.tab".format(i)))
tree = tree_learner(data)
pickle_path = os.path.join(results_path, "test-pickle.pkl")
pickle_obj(tree, pickle_path)
