def bernoulli_bayes_test(train_files_number=10, lines_per_file=10):
text_classifier = BayesTextClassifier()
train_classifier(text_classifier, train_files_number, resource_folder="resources")
case_count, correct_case_count = 0, 0
for expected_language, listing in get_test_set(10, lines_per_file):
actual_language = text_classifier.test(listing, model='bernoulli')
case_count += 1
if expected_language == actual_language:
correct_case_count += 1
print 'Bernoulli Bayes Test. Model was trained on %d files per language. %d lines per test file. ' \
'%d out of %d are correctly classifier. Productivity: %f' \
% (train_files_number, lines_per_file, correct_case_count, case_count, float(correct_case_count / case_count))
def multinomial_bayes_test(train_files_number=10, lines_per_file=10, feature_selection_method='mutual', feature_selection_count=-1):
text_classifier = BayesTextClassifier()
train_classifier(text_classifier, train_files_number, resource_folder="resources")
if feature_selection_count != -1:
if feature_selection_method == 'mutual':
text_classifier.make_feature_selection_mutual_information(feature_selection_count)
elif feature_selection_method == 'square':
text_classifier.make_feature_selection_chi_square(feature_selection_count)
case_count, correct_case_count = 0, 0
for expected_language, listing in get_test_set(10, lines_per_file):
actual_language = text_classifier.test(listing)
case_count += 1
if expected_language == actual_language:
correct_case_count += 1
print 'Multinomial Bayes Test. Model was trained on %d files per language. %d lines per test file. ' \
'%d out of %d are correctly classifier. Productivity: %f' \
% (train_files_number, lines_per_file, correct_case_count, case_count, float(correct_case_count / case_count))
