def _test_semi(NaiveBayes):
clf = MultinomialNB(alpha=.01)
clf.fit(X_train, y_train)
# clf.fit(X_train[:n], Y_train[:n])
pred = clf.predict(X_test)
f1_score = metrics.f1_score(y_test, pred)
print "Supervised Learning - scikits-learn 0.10"
# print "Accuracy: %d/%d" % (np.sum(pred == y_test), len(y_test))
accuracy = float( np.sum(pred == y_test) ) / float(len(y_test))
print "Accuracy: %0.4f" % accuracy
print "F1: %0.4f" % f1_score
clf = NaiveBayes()
clf.fit(X_train, y_train)
# clf.fit(X_train[:n], Y_train[:n])
pred = clf.predict(X_test)
f1_score = metrics.f1_score(y_test, pred)
print "Supervised Learning"
# print "Accuracy: %d/%d" % (np.sum(pred == y_test), len(y_test))
accuracy = float( np.sum(pred == y_test) ) / float(len(y_test))
print "Accuracy: %0.4f" % accuracy
print "F1: %0.4f" % f1_score
clf = NaiveBayes()
clf.fit_semi(X_train, Y_train, X_unlabel)
# clf.fit_semi(X_train[:n], Y_train[:n], X_train[n:])
pred = clf.predict(X_test)
f1_score = metrics.f1_score(y_test, pred)
accuracy = float( np.sum(pred == y_test) ) / float(len(y_test))
print "Semi-Supervised Learning"
print "Accuracy: %0.4f" % accuracy
print "F1: %0.4f" % f1_score
print "-----"
