def makeLearner(self):
# for icmi
#from orangeUtils import ThresholdProbabilityLearner
#import orngBayes
#learner = orngBayes.BayesLearner()
#learner.adjustThreshold = True
#return learner
#return orngEnsemble.RandomForestLearner(data)
#return orngTree.TreeLearner(data)
#return treefss.TreeFSS(N=7)(data)
treeLearner = orngTree.TreeLearner(storeExamples=True)
treeLearner.stop = orange.TreeStopCriteria_common()
#treeLearner.stop.minExamples = 1
treeLearner.maxDepth = 5
#treeLearner.stop.maxMajority = 0.8
#return treeLearner
return preposition.RejectInsaneExampleLearner(treeLearner)
def learnModel(self, X, y):
if numpy.unique(y).shape[0] != 2:
raise ValueError("Can only operate on binary data")
classes = numpy.unique(y)
self.worstResponse = classes[classes != self.bestResponse][0]
#We need to convert y into indices
newY = self.labelsToInds(y)
XY = numpy.c_[X, newY]
attrList = []
for i in range(X.shape[1]):
attrList.append(orange.FloatVariable("X" + str(i)))
attrList.append(orange.EnumVariable("y"))
attrList[-1].addValue(str(self.bestResponse))
attrList[-1].addValue(str(self.worstResponse))
self.domain = orange.Domain(attrList)
eTable = orange.ExampleTable(self.domain, XY)
#Weight examples and equalise
#Equalizing computes such weights that the weighted number of examples
#in each class is equivalent.
preprocessor = orange.Preprocessor_addClassWeight(equalize=1)
preprocessor.classWeights = [1 - self.weight, self.weight]
eTable, weightID = preprocessor(eTable)
eTable.domain.addmeta(weightID, orange.FloatVariable("w"))
self.learner = orngTree.TreeLearner(m_pruning=self.m,
measure="gainRatio")
self.learner.max_depth = self.maxDepth
self.learner.stop = orange.TreeStopCriteria_common()
self.learner.stop.min_instances = self.minSplit
self.classifier = self.learner(eTable, weightID)
def instance(self):
learner = orange.TreeLearner()
hasSplit = hasattr(self, "split")
if hasSplit:
learner.split = self.split
else:
learner.split = orange.TreeSplitConstructor_Combined()
learner.split.continuousSplitConstructor = orange.TreeSplitConstructor_Threshold(
)
binarization = getattr(self, "binarization", 0)
if binarization == 1:
learner.split.discreteSplitConstructor = orange.TreeSplitConstructor_ExhaustiveBinary(
)
elif binarization == 2:
learner.split.discreteSplitConstructor = orange.TreeSplitConstructor_OneAgainstOthers(
)
else:
learner.split.discreteSplitConstructor = orange.TreeSplitConstructor_Attribute(
)
measures = {
"infoGain": orange.MeasureAttribute_info,
"gainRatio": orange.MeasureAttribute_gainRatio,
"gini": orange.MeasureAttribute_gini,
"relief": orange.MeasureAttribute_relief,
"retis": orange.MeasureAttribute_MSE
}
measure = getattr(self, "measure", None)
if type(measure) == str:
measure = measures[measure]()
if not hasSplit and not measure:
measure = orange.MeasureAttribute_gainRatio()
measureIsRelief = type(measure) == orange.MeasureAttribute_relief
relM = getattr(self, "reliefM", None)
if relM and measureIsRelief:
measure.m = relM
relK = getattr(self, "reliefK", None)
if relK and measureIsRelief:
measure.k = relK
learner.split.continuousSplitConstructor.measure = measure
learner.split.discreteSplitConstructor.measure = measure
wa = getattr(self, "worstAcceptable", 0)
if wa:
learner.split.continuousSplitConstructor.worstAcceptable = wa
learner.split.discreteSplitConstructor.worstAcceptable = wa
ms = getattr(self, "minSubset", 0)
if ms:
learner.split.continuousSplitConstructor.minSubset = ms
learner.split.discreteSplitConstructor.minSubset = ms
if hasattr(self, "stop"):
learner.stop = self.stop
else:
learner.stop = orange.TreeStopCriteria_common()
mm = getattr(self, "maxMajority", 1.0)
if mm < 1.0:
learner.stop.maxMajority = self.maxMajority
me = getattr(self, "minExamples", 0)
if me:
learner.stop.minExamples = self.minExamples
for a in [
"storeDistributions", "storeContingencies", "storeExamples",
"storeNodeClassifier", "nodeLearner", "maxDepth"
]:
if hasattr(self, a):
setattr(learner, a, getattr(self, a))
return learner
def makeLearner(self):
#return orngEnsemble.RandomForestLearner()
treeLearner = orngTree.TreeLearner(storeExamples=True)
treeLearner.stop = orange.TreeStopCriteria_common()
treeLearner.maxDepth=4
return preposition.RejectInsaneExampleLearner(treeLearner)
nodeCont = node.distribution
majorClass = node.nodeClassifier.defaultValue
print "--> %s (%s) " % (majorClass, nodeCont),
def printTree(x):
if type(x) == orange.TreeClassifier:
printTree0(x.tree, 0)
elif type(x) == orange.TreeNode:
printTree0(x, 0)
else:
raise TypeError, "invalid parameter"
print learner.split
learner(data)
print learner.split
learner.stop = orange.TreeStopCriteria_common()
print learner.stop.maxMajority, learner.stop.minExamples
print "\n\nTree with minExamples = 5.0"
learner.stop.minExamples = 5.0
tree = learner(data)
printTree(tree)
print "\n\nTree with maxMajority = 0.5"
learner.stop.maxMajority = 0.5
tree = learner(data)
printTree(tree)
def makeLearner(self):
treeLearner = orngTree.TreeLearner(storeExamples=True)
treeLearner.stop = orange.TreeStopCriteria_common()
treeLearner.maxDepth = 5
return preposition.RejectInsaneExampleLearner(treeLearner)