def test1(self):
classifier = nbc.Bayes_Classifier()
classifier.train(data[:12478])
predictions = classifier.classify(data[12478:])
fp, fn = f_score(data[12478:], predictions)
print(fp, fn)
self.assertGreater(fp, 0.90)
self.assertGreater(fn, 0.60)
def main():
source_name = sys.argv[0]
check_imports(source_name)
bayes = nbc.Bayes_Classifier()
bayes.train('train.txt')
predict = bayes.classify('classifyA.txt')
fA = f_score('answersA.txt', predict)
print('Classifier F-Scores:')
print(' Reviews A: %.2f' % fA)
exit(0)
def main():
source_name = sys.argv[0]
check_imports(source_name)
bayes = nbc.Bayes_Classifier()
bayes.train('/home/xwl7110/EECS348_Labs/Labs/Lab5/src/train.txt')
predict = bayes.classify(
'/home/xwl7110/EECS348_Labs/Labs/Lab5/src/classifyA.txt')
fA = f_score('/home/xwl7110/EECS348_Labs/Labs/Lab5/src/answersA.txt',
predict)
print('Classifier F-Scores:')
print(predict)
print(' Reviews A: %.2f' % fA)
def test(self):
sum_fp = 0
sum_fn = 0
for i in range(1, 11):
k = i * 1400
classifier = nbc.Bayes_Classifier()
classifier.train(data[0:k - 1400] + data[k:])
predictions = classifier.classify(data[k - 1400:k])
fp, fn = f_score(data[k - 1400:k], predictions)
print(fp, fn)
self.assertGreater(fp, 0.90)
self.assertGreater(fn, 0.60)
sum_fp += fp
sum_fn += fn
print(" the value of K is {}".format(k))
print("the average FP is {} and the average FN is {}".format(
sum_fp / 10, sum_fn / 10))
fp = fp + 1
# print 'predicted 5 but actually 1', ids[i]
if predict[i] == '1' and actual[i] == '1':
tn = tn + 1
if predict[i] == '1' and actual[i] == '5':
fn = fn + 1
# print 'predicted 1 but actually 5', ids[i]
precision = float(tp) / float(tp + fp)
recall = float(tp) / float(tp + fn)
f_score = float(2.0) * precision * recall / (precision + recall)
return (f_score)
bc1 = nbc.Bayes_Classifier()
bc1.train('train_short.txt')
predict = bc1.classify('halfA.txt')
fA = f_score('answershalfA.txt', predict)
print 'less skew data f score is', fA
# bc2 = nbc.Bayes_Classifier()
# bc2.train_improve('train_short.txt')
# predict = bc2.classify_improve('halfA.txt')
# fA = f_score('answershalfA.txt',predict)
# print 'less skew improved? f is', fA
bc3 = nbc.Bayes_Classifier()
bc3.train('train.txt')
predict = bc3.classify('classifyA.txt')
fA = f_score('answersA.txt', predict)