def evaluate_and_print():
"""Evaluates the system and outputs per-network results to separate
CSV files."""
printer = pretty_print_summary
twitter_entries = get_twitter_entries()
twitter_start_time = int(time.time())
print 'Evaluating Twitter', twitter_start_time
twitter_evaluator = Evaluator(twitter_entries, TwitterProfile,
TwitterSearch(), 'twitter')
twitter_results = twitter_evaluator.evaluate()
twitter_end_time = int(time.time())
twitter_lines = ['\n' + x for x in printer('Twitter', twitter_results)]
print 'Done', twitter_end_time, '. Total',\
twitter_end_time - twitter_start_time
with open('twitter_evaluation_results.csv', 'w') as twitter_result_file:
twitter_result_file.writelines(twitter_lines)
facebook_entries = get_facebook_entries()
facebook_start_time = int(time.time())
print 'Evaluating Facebook', facebook_start_time
facebook_evaluator = Evaluator(facebook_entries, FacebookProfile,
FacebookSearch(), 'facebook')
facebook_results = facebook_evaluator.evaluate()
facebook_end_time = int(time.time())
facebook_lines = ['\n' + x for x in printer('Facebook', facebook_results)]
print 'Done', facebook_end_time, '. Total',\
facebook_end_time - facebook_start_time
with open('facebook_evaluation_results.csv', 'w') as facebook_result_file:
facebook_result_file.writelines(facebook_lines)
output_lines = []
output_lines += twitter_lines
output_lines += ['\n']
output_lines += facebook_lines
with open('evaluation_results.csv', 'w') as result_file:
result_file.writelines(output_lines)
def evaluate_to_file(converter=config.evaluate_converter, idx=0):
"""Evaluates the performance of the system, and saves the results
to data files in network-specific folders."""
name = converter.name
basedir = os.path.dirname(__file__)
results_directory = os.path.join(basedir, 'results')
twitter_entries = get_twitter_entries()
twitter_start_time = int(time.time())
print 'Evaluating Twitter', name, twitter_start_time
twitter_evaluator = Evaluator(twitter_entries, TwitterProfile,
TwitterSearch(), 'twitter',
converter=converter)
twitter_results = twitter_evaluator.evaluate()
twitter_end_time = int(time.time())
processed_results = process_network_results(twitter_results)
twitter_folder = os.path.join(results_directory, 'Twitter')
results_file = os.path.join(twitter_folder, '%02d. ' % idx + name)
cPickle.dump(processed_results, open(results_file, 'wb'))
print 'Done', twitter_end_time, '. Total',\
twitter_end_time - twitter_start_time
facebook_entries = get_facebook_entries()
facebook_start_time = int(time.time())
print 'Evaluating Facebook', facebook_start_time
facebook_evaluator = Evaluator(facebook_entries, FacebookProfile,
FacebookSearch(), 'facebook',
converter=converter)
facebook_results = facebook_evaluator.evaluate()
facebook_end_time = int(time.time())
processed_results = process_network_results(facebook_results)
facebook_folder = os.path.join(results_directory, 'Facebook')
results_file = os.path.join(facebook_folder, '%02d. ' % idx + name)
cPickle.dump(processed_results, open(results_file, 'wb'))
print 'Done', facebook_end_time, '. Total',\
facebook_end_time - facebook_start_time
def evaluate_statistical_significance(converter=config.evaluate_converter,
idx=0):
"""Evaluates the performance of the system in terms of number of correct
results, and saves the results to data files in network-specific folders.
Used to determine if feature sets show statistical significance."""
name = converter.__name__
basedir = os.path.dirname(__file__)
results_directory = os.path.join(basedir, 'results')
twitter_entries = get_twitter_entries()
twitter_start_time = int(time.time())
print 'Evaluating Twitter', name, twitter_start_time
twitter_evaluator = Evaluator(twitter_entries, TwitterProfile,
TwitterSearch(), 'twitter',
converter=converter)
twitter_results = twitter_evaluator.evaluate_statistical()
twitter_end_time = int(time.time())
twitter_folder = os.path.join(results_directory, 'Twitter')
results_file = os.path.join(twitter_folder, '%02d. ' % idx + name)
cPickle.dump(twitter_results, open(results_file, 'wb'))
print 'Done', twitter_end_time, '. Total',\
twitter_end_time - twitter_start_time
facebook_entries = get_facebook_entries()
facebook_start_time = int(time.time())
print 'Evaluating Facebook', facebook_start_time
facebook_evaluator = Evaluator(facebook_entries, FacebookProfile,
FacebookSearch(), 'facebook',
converter=converter)
facebook_results = facebook_evaluator.evaluate_statistical()
facebook_end_time = int(time.time())
facebook_folder = os.path.join(results_directory, 'Facebook')
results_file = os.path.join(facebook_folder, '%02d. ' % idx + name)
cPickle.dump(facebook_results, open(results_file, 'wb'))
print 'Done', facebook_end_time, '. Total',\
facebook_end_time - facebook_start_time