Exemple #1
0
from models.perceptron import Perceptron

model = Perceptron(train='pa2_train_clean.csv',
                   validation='pa2_valid_clean.csv',
                   test='pa2_test_no_label_clean.csv',
                   label='label',
                   mod_type='average',
                   max_iter=15)

learned_model = model.train_model()

# Save output for learned model to .json file.
output_folder = Path('model_output')
output_path = Path(__file__).parent.resolve().joinpath(output_folder)
training_file = output_path.joinpath(Path('ap_training.json'))

# Create output directory if doesn't exist.
if not Path(output_path).exists():
    Path(output_path).mkdir()
with open(training_file, 'w') as f:
    json.dump(learned_model, f, indent=4)

# Best validation accuracy with 15 iterations, calculate and save predictions.
test_predictions = model.predict_test(learned_model['avg_weights'][-1])
prediction_file = output_path.joinpath(Path('aplabel.csv'))
with open(prediction_file, 'w') as fp:
    writer = csv.writer(fp)
    for i in test_predictions:
        writer.writerows([[i]])
Exemple #2
0
from models.perceptron import Perceptron

model = Perceptron(train='pa2_train_clean.csv',
                   validation='pa2_valid_clean.csv',
                   test='pa2_test_no_label_clean.csv',
                   label='label',
                   mod_type='online',
                   max_iter=15)

learned_model = model.train_model()

# Save output for learned model to .json file.
output_folder = Path('model_output')
output_path = Path(__file__).parent.resolve().joinpath(output_folder)
training_file = output_path.joinpath(Path('op_training.json'))

# Create output directory if doesn't exist.
if not Path(output_path).exists():
    Path(output_path).mkdir()
with open(training_file, 'w') as f:
    json.dump(learned_model, f, indent=4)

# Best validation accuracy with 14 iterations, calculate and save predictions.
test_predictions = model.predict_test(learned_model['weights'][13])
prediction_file = output_path.joinpath(Path('oplabel.csv'))
with open(prediction_file, 'w') as fp:
    writer = csv.writer(fp)
    for i in test_predictions:
        writer.writerows([[i]])