def xval(start,stop,data,rows,f,z,k,m,abcd):
rmax = len(rows)
test = []
hypotheses = {}
temp = ""
for r in range(0, rmax):
d = rows[r]
if r >= start and r < stop:
test.append(d)
else:
temp = klass1(d, z)
try:
hypotheses[temp] += 1
if hypotheses[temp] == 1:
makeTable(colname[z],temp)
addRow(d,temp)
except KeyError:
hypotheses[temp] = 1
if hypotheses[temp] == 1:
makeTable(colname[z],temp)
addRow(d,temp)
#zeror(test, data, hypotheses, z)
#xvalTest1(test,data,hypotheses)
#return nb(test,data,hypotheses,z,k,m,abcd)
nb(test,data,hypotheses,z,k,m,abcd)
def xval(start, stop, data, rows, f, z, k, m, abcd):
rmax = len(rows)
test = []
hypotheses = {}
temp = ""
for r in range(0, rmax):
d = rows[r]
if r >= start and r < stop:
test.append(d)
else:
temp = klass1(d, z)
try:
hypotheses[temp] += 1
if hypotheses[temp] == 1:
makeTable(colname[z], temp)
addRow(d, temp)
except KeyError:
hypotheses[temp] = 1
if hypotheses[temp] == 1:
makeTable(colname[z], temp)
addRow(d, temp)
#zeror(test, data, hypotheses, z)
#xvalTest1(test,data,hypotheses)
#return nb(test,data,hypotheses,z,k,m,abcd)
nb(test, data, hypotheses, z, k, m, abcd)
def _setup_lion_classifier(self):
classifier = nb(feature_labels, class_label)
classifier._class_node._truth_table = numpy.array([[0.9, 0.1]],
dtype=float)
classifier._feature_nodes["has-fur?"]._truth_table = numpy.array(
[[0.9, 0.1], [0.5, 0.5]], dtype=float)
classifier._feature_nodes["long-teeth?"]._truth_table = numpy.array(
[[0.5, 0.5], [0.1, 0.9]], dtype=float)
classifier._feature_nodes["scary?"]._truth_table = numpy.array(
[[0.5, 0.5], [0.2, 0.8]], dtype=float)
return classifier
def test_train(self):
classifier = nb(feature_labels, class_label)
dataset = [
{
"has-fur?": True,
"long-teeth?": False,
"scary?": False
},
{
"has-fur?": False,
"long-teeth?": True,
"scary?": True
},
{
"has-fur?": True,
"long-teeth?": True,
"scary?": True
},
]
annotations = [
False,
False,
True,
]
classifier.train(dataset, annotations)
lion_node = classifier._class_node
has_fur_node = classifier._feature_nodes["has-fur?"]
long_teeth_node = classifier._feature_nodes["long-teeth?"]
scary_node = classifier._feature_nodes["scary?"]
conditions_table = numpy.array([[0.5, 0.5], [0.0, 1.0]], dtype=float)
self.assertTrue(
numpy.array_equal(
numpy.array([[0.67, 0.33]]),
numpy.array(
[[numpy.round(x, 2) for x in lion_node._truth_table[0]]])))
self.assertTrue(
numpy.array_equal(
conditions_table,
numpy.array([[numpy.round(y, 2) for y in x]
for x in has_fur_node._truth_table])))
self.assertTrue(
numpy.array_equal(
conditions_table,
numpy.array([[numpy.round(y, 2) for y in x]
for x in long_teeth_node._truth_table])))
self.assertTrue(
numpy.array_equal(
conditions_table,
numpy.array([[numpy.round(y, 2) for y in x]
for x in scary_node._truth_table])))
#Genre!
else:
trainFeaturesAndLabels = [(featureExtractor(lyrics, words), genre) for (lyrics, genre) in trainSongs]
testFeaturesAndLabels = [(featureExtractor(lyrics, words), genre) for (lyrics, genre) in testSongs]
thisTime = time.clock()
print "Extract features: ", thisTime - lastTime, ' s'
lastTime = thisTime
#Train classifier
nbClassifier = nbTrain(trainFeaturesAndLabels, labels, words)
thisTime = time.clock()
print "Train Classifier: ", thisTime - lastTime, ' s'
lastTime = thisTime
#Test for errors
trainError = nbClassifier.getErrorRate(trainFeaturesAndLabels)
testError = nbClassifier.getErrorRate(testFeaturesAndLabels)
thisTime = time.clock()
print "Error checking: ", thisTime - lastTime, ' s'
lastTime = thisTime
print "Train Error: ", trainError
print "Test Error: ", testError, " with", len(labels), "labels"
if __name__ == "__main__":
nb()
def test_constructor(self):
classifier = nb(feature_labels, class_label)
self.assertEqual(len(classifier._feature_nodes), 3)
self.assertEqual(classifier._class_node._label.get_label(), "lion")
for feature_name in ["has-fur?", "long-teeth?", "scary?"]:
self.assertTrue(feature_name in classifier._feature_nodes.keys())
def getRandomBinaryChromosome(n): ans=[] for i in range(n): ans.append(choice(["1","0"])) return ans seed(0) done=False limit=100 conv_limit=0.1 counter=0 oldfitness=[] pop_size=10 crossover_rate=0.3 mutation_rate=0.25 n=nb(0) # Initialize population randomly population=[getRandomBinaryChromosome(n) for i in range(pop_size)] while not done and counter<limit: counter+=1 #print population done=True # Fitness evaluation fitness=[nb(None,i) for i in population] if oldfitness!=[]: for i in range(pop_size): if abs(fitness[i]-oldfitness[i])>conv_limit: done=False if done: print "Fitness converged! Iteration %d" % counter break
def _nb1():
with settings(LIB,seed=2),settings(TABLE,era=3):
nb("weather.csv")
def _nb1():
with settings(LIB,seed=1),settings(TABLE,era=50):
nb("housingD.csv")
trainFeaturesAndLabels = [(featureExtractor(lyrics, words), genre)
for (lyrics, genre) in trainSongs]
testFeaturesAndLabels = [(featureExtractor(lyrics, words), genre)
for (lyrics, genre) in testSongs]
thisTime = time.clock()
print "Extract features: ", thisTime - lastTime, ' s'
lastTime = thisTime
#Train classifier
nbClassifier = nbTrain(trainFeaturesAndLabels, labels, words)
thisTime = time.clock()
print "Train Classifier: ", thisTime - lastTime, ' s'
lastTime = thisTime
#Test for errors
trainError = nbClassifier.getErrorRate(trainFeaturesAndLabels)
testError = nbClassifier.getErrorRate(testFeaturesAndLabels)
thisTime = time.clock()
print "Error checking: ", thisTime - lastTime, ' s'
lastTime = thisTime
print "Train Error: ", trainError
print "Test Error: ", testError, " with", len(labels), "labels"
if __name__ == "__main__":
nb()
