while not eof:
line = f_pos.readline()
eof = not len(line)
x.append(count_words(line.decode('utf-8', 'ignore')))
y.append(+1)
x.append(count_words(f_neg.readline().decode('utf-8', 'ignore')))
y.append(-1)
i += 1
fold = int(i * 0.9)
f_pos.close()
f_neg.close()
from ml.nb import NaiveBayes
classifier = NaiveBayes()
classifier.train(x[0:fold], y[0:fold])
estim_y = classifier.predict(x[fold:])
(acc, ) = classifier.evaluate(y[fold:], estim_y)
print('Naive Bayes accuracy = {0:.2f}%'.format(acc * 100))
from ml.svm import SVM
classifier = SVM()
classifier.train(x[0:fold], y[0:fold])
estim_y = classifier.predict(x[fold:])
(acc, ) = classifier.evaluate(y[fold:], estim_y)
print('SVM accuracy = {0:.2f}%'.format(acc * 100))
# divide into training and test sets
test_x = []
test_y = []
train_x = []
train_y = []
for i in range(0, len(x)):
if i % 3:
train_x.append(list_to_dict(x[i]))
train_y.append(y[i])
else:
test_x.append(list_to_dict(x[i]))
test_y.append(y[i])
from ml.nb import NaiveBayes
classifier = NaiveBayes()
classifier.train(train_x, train_y)
estim_y = classifier.predict(test_x)
(acc, ) = classifier.evaluate(test_y, estim_y)
print('Naive Bayes accuracy = {0:.2f}%'.format(acc * 100))
from ml.svm import SVM
classifier = SVM()
classifier.train(train_x, train_y)
estim_y = classifier.predict(test_x)
(acc, ) = classifier.evaluate(test_y, estim_y)
print('SVM accuracy = {0:.2f}%'.format(acc * 100))
while not eof:
line = f_pos.readline()
eof = not len(line)
x.append(count_words(line.decode('utf-8', 'ignore')))
y.append(+1)
x.append(count_words(f_neg.readline().decode('utf-8', 'ignore')))
y.append(-1)
i += 1
fold = int(i * 0.9)
f_pos.close()
f_neg.close()
from ml.nb import NaiveBayes
classifier = NaiveBayes()
classifier.train(x[0:fold], y[0:fold])
estim_y = classifier.predict(x[fold:])
(acc, ) = classifier.evaluate(y[fold:], estim_y)
print('Naive Bayes accuracy = {0:.2f}%'.format(acc * 100))
from ml.svm import SVM
classifier = SVM()
classifier.train(x[0:fold], y[0:fold])
estim_y = classifier.predict(x[fold:])
(acc, ) = classifier.evaluate(y[fold:], estim_y)
print('SVM accuracy = {0:.2f}%'.format(acc * 100))
def list_to_dict(l):
# return dict(zip(l, [1 for _ in range(0, len(l))]))
return dict(zip(range(0, len(l)), l))
# divide into training and test sets
test_x = []; test_y = []; train_x = []; train_y = [];
for i in range(0, len(x)):
if i % 3:
train_x.append(list_to_dict(x[i]))
train_y.append(y[i])
else:
test_x.append(list_to_dict(x[i]))
test_y.append(y[i])
from ml.nb import NaiveBayes
classifier = NaiveBayes()
classifier.train(train_x, train_y)
estim_y = classifier.predict(test_x)
(acc, ) = classifier.evaluate(test_y, estim_y)
print('Naive Bayes accuracy = {0:.2f}%'.format(acc * 100))
from ml.svm import SVM
classifier = SVM()
classifier.train(train_x, train_y)
estim_y = classifier.predict(test_x)
(acc, ) = classifier.evaluate(test_y, estim_y)
print('SVM accuracy = {0:.2f}%'.format(acc * 100))
(name, sex) = line.strip().split(' ')
x.append({name[-1:]: 1, name[-2:]: 1, name[-3:]: 1, name[-4:]: 1})
y.append(sex)
fold = -100
f.close()
import os
from ml.nb import NaiveBayes
filename = 'tmp/nb.pickle'
exists = os.path.isfile(filename)
if exists:
classifier = NaiveBayes.load(filename)
else:
classifier = NaiveBayes()
classifier.train(x[0:fold], y[0:fold])
estim_y = classifier.predict(x[fold:])
(acc, ) = classifier.evaluate(y[fold:], estim_y)
if not exists:
classifier.save(filename)
print('Naive Bayes accuracy = {0:.2f}%'.format(acc * 100))
from ml.svm import SVM
filename = 'tmp/svm.pickle'
