def get_avg_maps(train_data):
chunk_avg = utilities.get_chunk_avg(train_data)
hour_avg = utilities.get_hour_avg(train_data)
hour_avg_by_chunk = utilities.get_hour_avg_by_chunk(train_data)
weekday_avg = utilities.get_weekday_avg(train_data)
weekday_avg_by_chunk = utilities.get_weekday_avg_by_chunk(train_data)
return (chunk_avg, hour_avg_by_chunk, weekday_avg_by_chunk,
hour_avg, weekday_avg)
def baseline(training_file, submission_file, output_file):
data = utilities.read_file(training_file)
sub_data = utilities.read_file(submission_file, True)
print 'Calculating hour averages...'
hour_avg_by_chunk = utilities.get_hour_avg_by_chunk(data)
hour_avg = utilities.get_hour_avg(data)
print 'Filling submission file...'
for i in range(1, len(sub_data)):
chunk_id = sub_data[i][1]
hour = sub_data[i][3]
for j in range(5, len(sub_data[i])):
if sub_data[i][j] == '0':
if chunk_id in hour_avg_by_chunk:
sub_data[i][j] = hour_avg_by_chunk[chunk_id][hour][j - 5]
else:
sub_data[i][j] = hour_avg[hour][j - 5]
utilities.write_file(output_file, sub_data)
def baseline(training_file, submission_file, output_file):
data = utilities.read_file(training_file)
sub_data = utilities.read_file(submission_file, True)
print 'Calculating hour averages...'
hour_avg_by_chunk = utilities.get_hour_avg_by_chunk(data)
hour_avg = utilities.get_hour_avg(data)
print 'Filling submission file...'
for i in range(1, len(sub_data)):
chunk_id = sub_data[i][1]
hour = sub_data[i][3]
for j in range(5, len(sub_data[i])):
if sub_data[i][j] == '0':
if chunk_id in hour_avg_by_chunk:
sub_data[i][j] = hour_avg_by_chunk[chunk_id][hour][j - 5]
else:
sub_data[i][j] = hour_avg[hour][j - 5]
utilities.write_file(output_file, sub_data)
