def train_and_save(classifier_list, classifier_name_list, training_data):
results = train_multiple_classifiers(classifier_list, classifier_name_list,
training_data)
util.save_object(results, RESULTS_PATH)
util.save_classifier_list(classifier_list, classifier_name_list,
CLASSIFIERS_AND_RESULTS_DIR_PATH)
return results
def train_and_save(classifier_list, classifier_name_list, training_data):
global results
results.append(
train_multiple_classifiers(classifier_list, classifier_name_list,
training_data))
util.save_object(results, RESULTS_PATH)
util.save_classifier_list(classifier_list, classifier_name_list,
CLASSIFIERS_AND_RESULTS_DIR_PATH)
def run_experiment(self):
print('Running', self.exp_name, ', it.', self.classifier_iter)
# Load dataset
self.data_df = self.get_dataset_from_name(self.dataset_enum)
self.avg_dataset_length = get_dataset_avg_length(self.data_df)
print('Got dataset:', self.dataset_enum, 'num of cat.',
self.get_num_of_categories())
# Split on train and test dataset
self.train_corpus, self.test_corpus, self.train_label_names, \
self.test_label_names = \
train_test_split(np.array(self.data_df['Clean Article']),
np.array(self.data_df['Target Name']),
test_size=self.TEST_SET_SIZE_RATIO, random_state=42)
# Tokenize corpus
self.tokenized_train = [
tn.tokenizer.tokenize(text) for text in self.train_corpus
]
self.tokenized_test = [
tn.tokenizer.tokenize(text) for text in self.test_corpus
]
# Get list of words (I know, cool, not readable one-liner)
data_word_list = ''.join(list(
self.data_df['Clean Article'])).split(' ')
self.vocabulary = set(data_word_list)
# Calculate features
self.train_features, self.test_features = self.get_features()
print('Train features shape:', self.train_features.shape,
' Test features shape:', self.test_features.shape)
# Pack data in one class
self.training_data = TrainingData(self.train_features,
self.train_label_names,
self.test_features,
self.test_label_names)
# Perform actual training
self.results = train_multiple_classifiers(
self.classifier_list, self.classifier_name_list,
self.training_data, self.CLASSIFIERS_AND_RESULTS_DIR_PATH,
self.classifier_iter, self.RESULTS_PATH)
# Extract scores for plotting
self.display_results()
pass
np.array(data_df['Target Name']),
test_size=TEST_SET_SIZE_RATIO,
random_state=42)
cv_train_features, cv_test_features = get_simple_bag_of_words_features(
train_corpus, test_corpus)
# # pack data in one class
training_data = TrainingData(cv_train_features, train_label_names,
cv_test_features, test_label_names)
# # Get classifier definitions
classifier_list, classifier_name_list, classifier_name_shortcut_list = \
get_chosen_classifiers()
# Train and save on disk
results = train_multiple_classifiers(classifier_list, classifier_name_list,
training_data)
# # Load from disk
# classifier_list = util.load_classifier_list(classifier_name_list,
# CLASSIFIERS_AND_RESULTS_DIR_PATH)
# results = util.load_object(RESULTS_PATH)
# create_cv_test_time_plots(results, classifier_name_shortcut_list)
cv_mean_scores = [round(result[1], SCORE_DECIMAL_PLACES) for result in results]
test_scores = [round(result[2], SCORE_DECIMAL_PLACES) for result in results]
elapsed_times = [round(result[3], TIME_DECIMAL_PLACES) for result in results]
# create_bar_plot(classifier_name_shortcut_list, 'Classifier scores', 'Accuracy',
# cv_mean_scores, y_range_tuple=(0, 1))
create_2_bar_plot(classifier_name_shortcut_list,
'Classifier scores',
'Accuracy',
cv_mean_scores,
