def __init__(self, path, bTrain):
self.bTrain = bTrain
if bTrain:
self.file = open(path+'.tab', "w")
else:
data = orange.ExampleTable(path)
self.file = orange.C45Learner(data, prune=False)
self.names = None
self.types = None
self.classes = None
self.examples = []
def train_classifier(data, type, filter):
if type == "tree" or type == "c4.5" or type == "decision_tree":
learner = orange.C45Learner()
elif type == "bayes" or type == "naive" or type == "naive_bayes":
learner = orange.BayesLearner()
elif type == "svm" or type == "linear_svm":
learner = Orange.classification.svm.LinearSVMLearner()
#elif type == "logreg" or type == "regression":
# learner = Orange.classification.logreg.LogRegLearner()
else:
print "Invalid Learner Type\n"
exit()
if filter == 0:
classifier = learner(data)
else:
filtered_learner = Orange.feature.selection.FilteredLearner(
learner,
filter=Orange.feature.selection.FilterBestN(n=filter),
name='filtered')
classifier = filtered_learner(data)
return classifier
# Description: Shows how to use C4.5 learner
# Category: learning
# Classes: C45Learner, C45Classifier
# Uses: iris
# Referenced: C45Learner.htm
import orange
#data = orange.ExampleTable("lenses.tab")
data = orange.ExampleTable("iris")
tree = orange.C45Learner(data)
print "\n\nC4.5 with default arguments"
for i in data[:5]:
print tree(i), i.getclass()
print "\n\nC4.5 with m=100"
tree = orange.C45Learner(data, m=100)
for i in data[:5]:
print tree(i), i.getclass()
print "\n\nC4.5 with minObjs=100"
tree = orange.C45Learner(data, minObjs=100)
for i in data[:5]:
print tree(i), i.getclass()
print "\n\nC4.5 with -m 1 and -s"
lrn = orange.C45Learner()
lrn.commandline("-m 1 -s")
tree = lrn(data)
for i in data:
def c45_tree(input_dict):
import orange
output_dict = {}
output_dict['c45out'] = orange.C45Learner(name="C4.5 Tree (Orange)")
return output_dict
def test_pickling_on(self, dataset):
try:
orange.C45Learner()
except orange.KernelException:
raise unittest.SkipTest("C45 dll not found")
testing.LearnerTestCase.test_pickling_on(self, dataset)
##f.close() ##print tree.name ##treeC45 = orange.C45Learner(trainD, minObjs=5) ##treeC45.name = "tree - C45 " ##f = file(trainData+'o.txt.C45tree', 'w') ##dumpC45Tree(treeC45,f) ##f.close() ##print treeC45.name t = orngTree.TreeLearner(measure='gainRatio', binarization=0, minSubset=2, minExamples=2, sameMajorityPruning=1, mForPruning=2); bsTree = BLearner(learner=t, examples=trainD) bsTree.name = "bsTree " print bsTree.name c45 = orange.C45Learner(minObjs=2) bsC45 = BLearner(learner=c45, examples=trainD) bsC45.name = "bsC45 " print bsC45.name ##svm = orange.SVMLearner(trainD) ##svm.name = "SVM " ##classifiers = [majority, tree, bsTree, treeC45, bsC45] classifiers = [majority, bsTree, bsC45] # compute accuracies print '='*80 acc = accuracy(trainD, classifiers) print "Classification accuracies - train:" print '-'*80