def run_part_8_code(data):
columns = [f'col_{num}' for num in range(8)]
index = [f'index_{num}' for num in range(data.shape[0])]
df = pd.DataFrame(data, columns=columns, index=index)
df.columns = [
'pollution', 'dew', 'temp', 'pressure', 'wind_dir', 'wind_spd', 'snow',
'rain'
]
df = df.drop(columns=['dew', 'temp', 'pressure', 'snow', 'rain'])
data_after_removal = df.values
train_X, train_y, test_X, test_y = data_processor.split_and_prep_data(
data_after_removal,
split_point=(constants.trian_split_part, constants.test_split_part),
window_size=constants.window_size,
method_name='generator')
# LSTM Model
lstm_1 = model.LSTM_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q8_lstm_model')
lstm_1.train_and_report_results()
# GRU Model
gru_1 = model.GRU_Model(train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q8_GRU_model')
gru_1.train_and_report_results()
# RNN Model
rnn_1 = model.RNN_Model(train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q8_RNN_model')
rnn_1.train_and_report_results()
def run_part_4_codes(train_X, train_y, test_X, test_y, time_Series_type):
# LSTM Model
lstm_1 = model.LSTM_Model(train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/' +
time_Series_type + 'Q4_lstm_model')
lstm_1.train_and_report_results()
# GRU Model
gru_1 = model.GRU_Model(train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/' +
time_Series_type + 'Q4_GRU_model')
gru_1.train_and_report_results()
# RNN Model
rnn_1 = model.RNN_Model(train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/' +
time_Series_type + 'Q4_RNN_model')
rnn_1.train_and_report_results()
def part_two_codes(data):
#part 1
train_data = data[:constants.trian_split_part]
other_data = data[constants.trian_split_part + 1:]
missed_data = missing_data.erase_data(train_data, 0.2)
# part 4
our_pred = missing_data.predict_method1(missed_data)
knn_pred = missing_data.predict_method2(missed_data)
mean_pred = missing_data.predict_method3(missed_data)
# part 5
evaluate_all_predictions(train_data, missed_data, our_pred, knn_pred,
mean_pred)
# part 6
new_data = np.concatenate((our_pred, other_data))
train_X, train_y, test_X, test_y = data_processor.split_and_prep_data(
new_data,
split_point=(constants.trian_split_part, constants.test_split_part),
window_size=constants.window_size,
method_name='generator')
print("LSTM")
lstm_1 = model.LSTM_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name=
'../Generated-Files/2Q6_generator_lstm_model_with_new_data')
lstm_1.train_and_report_results()
print("GRU")
gru_1 = model.GRU_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name=
'../Generated-Files/2Q6_generator_GRU_model_with_new_data')
gru_1.train_and_report_results()
train_X, train_y, test_X, test_y = data_processor.split_and_prep_data(
data,
split_point=(constants.trian_split_part, constants.test_split_part),
window_size=constants.window_size,
method_name='generator')
print("LSTM old data")
lstm_1 = model.LSTM_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name=
'../Generated-Files/2Q6_generator_lstm_model_with_old_data')
lstm_1.train_and_report_results()
print("GRU old data")
gru_1 = model.GRU_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name=
'../Generated-Files/2Q6_generator_GRU_model_with_old_data')
gru_1.train_and_report_results()
def run_window_11_codes_using_generator_data(train_X, train_y, test_X, test_y):
# LSTM Model
lstm_1 = model.LSTM_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q2_generator_lstm_model')
lstm_1.test_different_loss_and_optimization_functions()
lstm_1_with_dropout = model.LSTM_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=True,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q5_generator_lstm_with_dropout_model'
)
lstm_1_with_dropout.train_and_report_results()
# GRU Model
gru_1 = model.GRU_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q2_generator_GRU_model')
gru_1.test_different_loss_and_optimization_functions()
gru_1_with_dropout = model.GRU_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=True,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q5_generator_GRU_with_dropout_model'
)
gru_1_with_dropout.train_and_report_results()
# RNN Model
rnn_1 = model.RNN_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=constants.add_dropout,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q2_generator_RNN_model')
rnn_1.test_different_loss_and_optimization_functions()
rnn_1_with_dropout = model.RNN_Model(
train_X,
train_y,
test_X,
test_y,
loss_function=constants.loss_function,
optimizer=constants.optimizer,
validation_split=constants.validation_split,
add_dropout=True,
number_of_epochs=constants.number_of_epochs,
batch_size=constants.batch_size,
file_save_name='../Generated-Files/Q5_generator_RNN_with_dropout_model'
)
rnn_1_with_dropout.train_and_report_results()