def run_linear_regression(down_station, input_list, include_time, sample_size,
network_type, nr_layers, nr_units):
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
(y_train, x_train, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _,
_, _, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
#poly = PolynomialFeatures(degree=2)
#x_train_poly = poly.fit_transform(x_train)
#x_cv_poly = poly.fit_transform(x_cv)
#regr = linear_model.ElasticNet(alpha=1e-3, tol=1e-9)
#regr.fit(x_train, y_train)
#y_pred = regr.predict(x_cv)
# regr = linear_model.LinearRegression(fit_intercept=False)
regr = linear_model.LinearRegression()
regr.fit(x_train, y_train)
y_pred = regr.predict(x_cv)
util.save_linreg(regr, result_dir)
util.save_sklearn_model(regr, result_dir)
predict.plot_prediction(y_pred, result_dir, y_cv, train_y_max, train_y_min)
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Fitting time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run_gpr(down_station, input_list, include_time, sample_size, network_type,
include_diff, n_estimators, b):
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, n_estimators,
b, sample_size)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
(y_train, x_train, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _,
_, _, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
# n_estimators = 50
gpr = BaggingRegressor(GaussianProcessRegressor(copy_X_train=False),
max_samples=1.0 / n_estimators,
n_estimators=n_estimators,
n_jobs=1)
# svr = SVR(C=_C, epsilon=_epsilon, verbose=True, cache_size=1024) # No bagging
# gpr = GaussianProcessRegressor(copy_X_train=False)
gpr.fit(x_train, y_train)
util.save_sklearn_model(gpr, result_dir)
y_cv_pred = gpr.predict(x_cv)
predict.plot_prediction(y_cv_pred, result_dir, y_cv, train_y_max,
train_y_min)
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Fitting time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run_prob(down_station, input_list, include_time, sample_size, network_type,
nr_layers, nr_units):
"""Runner"""
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
(y_train, x_train, y_cv, x_cv, _, _, _, _, _, _, _, _, _, _,
_) = data.construct(down_station, input_list, include_time, sample_size,
network_type)
input_dim = 0
input_dim_2 = 0
if (network_type == 'bnn'):
(_, input_dim, input_dim_2) = x_train.shape
else:
(_, input_dim) = x_train.shape
my_model = model.create(result_dir, input_dim, nr_layers, nr_units,
network_type, input_dim_2)
trained_model_file = util.model_file_name_lowest_cv(result_dir)
my_model.load_weights(trained_model_file, by_name=True)
print(my_model.get_config())
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Training time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run_sklearn(down_station, input_list, include_time, sample_size,
network_type, nr_layers, nr_units):
"""Runner"""
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
(_, _, y_cv, x_cv, y_test, x_test, _, _, train_y_max, train_y_min, _, _, _,
_, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
my_model = util.load_sklearn_model(result_dir)
start_time = time.time()
# y_pred = my_model.predict(x_cv) # uncomment for validation
y_pred = my_model.predict(x_test) # uncomment for test
elapsed_time = time.time() - start_time
print(
time.strftime("Prediction done : %H:%M:%S", time.gmtime(elapsed_time)))
print("{0} predictions in {1} seconds ({2} per second)".format(
len(y_pred), elapsed_time,
len(y_pred) / elapsed_time))
# plot_prediction(y_pred, result_dir, y_cv, train_y_max, train_y_min) # uncomment for validation
plot_prediction(y_pred, result_dir, y_test, train_y_max,
train_y_min) # uncomment for test
def run(down_station, input_list, include_time, sample_size, network_type,
nr_layers, nr_units):
"""Runner"""
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
util.plot_training_performance(result_dir)
model_file = util.model_file_name(result_dir)
# model_file = util.model_file_name_lowest_cv(result_dir) # lowest cv model
my_model = util.load_model(model_file)
# uncomment for DWS prediction
# for specific dates, see internals of data.construct
#(_, _, _, _, _, _, _, _, train_y_max, train_y_min, _, _, _, full_x, full_y) = data.construct(down_station, input_list, include_time, sample_size, network_type)
#predict(my_model, result_dir, full_x, full_y, train_y_max, train_y_min)
# uncomment for normal prediction
#(_, _, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _, _, full_x, full_y) = data.construct(down_station, input_list, include_time, sample_size, network_type)
#predict(my_model, result_dir, x_cv, y_cv, train_y_max, train_y_min)
# uncomment for test prediction
(_, _, _, _, y_test, x_test, _, _, train_y_max, train_y_min, _, _, _,
full_x, full_y) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
predict(my_model, result_dir, x_test, y_test, train_y_max, train_y_min)
def produce_diff(down_station, input_list, include_time, sample_size,
network_type, _C, _epsilon, denormalise, roundit):
result_dir = util.get_result_dir(down_station, network_type, _C, _epsilon,
sample_size)
my_model = util.load_sklearn_model(result_dir)
data.produce_diff(my_model, result_dir, down_station, input_list,
include_time, sample_size, network_type, denormalise,
roundit)
def produce_diff(down_station, input_list, include_time, sample_size,
network_type, nr_layers, nr_units, denormalise, roundit):
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
model_file = util.model_file_name(result_dir)
my_model = util.load_model(model_file)
data.produce_diff(my_model, result_dir, down_station, input_list,
include_time, sample_size, network_type, denormalise,
roundit)
def run_svr(down_station, input_list, include_time, sample_size, network_type,
include_diff, _C, _epsilon):
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, _C, _epsilon,
sample_size)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
(y_train, x_train, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _,
_, _, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
#svr = GridSearchCV(SVR(kernel='rbf', gamma=0.1), cv=5,
# param_grid={"C": [1e0, 1e1, 1e2, 1e3],
# "gamma": np.logspace(-2, 2, 5)})
if network_type == 'svr':
n_estimators = 12
svr = BaggingRegressor(SVR(C=_C,
epsilon=_epsilon,
verbose=True,
cache_size=768),
max_samples=1.0 / n_estimators,
n_estimators=n_estimators,
n_jobs=-1)
# svr = SVR(C=_C, epsilon=_epsilon, verbose=True, cache_size=1024) # No bagging
else:
svr = LinearSVR(C=_C, epsilon=_epsilon, verbose=1, max_iter=20000)
svr.fit(x_train, y_train)
util.save_sklearn_model(svr, result_dir)
y_cv_pred = svr.predict(x_cv)
predict.plot_prediction(y_cv_pred, result_dir, y_cv, train_y_max,
train_y_min)
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Fitting time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run(down_station, input_list, include_time, sample_size, network_type,
nr_layers, nr_units, nr_epochs):
"""Runner"""
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
# down_station, input_list, include_time, sample_size, network_type
(y_train, x_train, y_cv, x_cv, _, _, _, _, _, _, _, _, _, _,
_) = data.construct(down_station, input_list, include_time, sample_size,
network_type)
input_dim = 0
input_dim_2 = 0
if (network_type == 'bnn') or (network_type == 'cnn') or (
network_type == 'rnn_lstm') or (network_type == 'rnn_gru'):
(_, input_dim, input_dim_2) = x_train.shape
elif (network_type == 'multi_cnn'):
input_dim = []
for x_train_i in x_train:
(_, input_dim_i, _) = x_train_i.shape
input_dim.append(input_dim_i)
else:
(_, input_dim) = x_train.shape
my_model = model.create(result_dir, input_dim, nr_layers, nr_units,
network_type, input_dim_2)
train(my_model, result_dir, y_train, x_train, y_cv, x_cv, nr_epochs)
util.plot_training_performance(result_dir)
predict.run(down_station, input_list, include_time, sample_size,
network_type, nr_layers, nr_units)
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Training time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run_sgdreg(down_station, input_list, include_time, sample_size,
network_type, _tol, _eta0):
start_time_run = time.time()
result_dir = util.get_result_dir(down_station, network_type, _tol, _eta0,
sample_size)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
(y_train, x_train, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _,
_, _, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
sgdreg = linear_model.SGDRegressor(max_iter=100000, tol=_tol, eta0=_eta0)
sgdreg.fit(x_train, y_train)
y_pred = sgdreg.predict(x_cv)
predict.plot_prediction(y_pred, result_dir, y_cv, train_y_max, train_y_min)
elapsed_time_run = time.time() - start_time_run
print(
time.strftime("Fitting time : %H:%M:%S",
time.gmtime(elapsed_time_run)))
def run_prob(down_station, input_list, include_time, sample_size, network_type,
nr_layers, nr_units):
"""Runner"""
result_dir = util.get_result_dir(down_station, network_type, nr_layers,
nr_units, sample_size)
(y_train, x_train, y_cv, x_cv, _, _, _, _, train_y_max, train_y_min, _, _,
_, _, _) = data.construct(down_station, input_list, include_time,
sample_size, network_type)
input_dim = 0
input_dim_2 = 0
if (network_type == 'bnn'):
(_, input_dim, input_dim_2) = x_train.shape
else:
(_, input_dim) = x_train.shape
my_model = model.create(result_dir, input_dim, nr_layers, nr_units,
network_type, input_dim_2)
trained_model_file = util.model_file_name_lowest_cv(result_dir)
my_model.load_weights(trained_model_file, by_name=True)
print(my_model.get_config())
predict(my_model, result_dir, x_cv, y_cv, train_y_max, train_y_min)