def standard_all(model_,
model_name,
big_X_train,
big_y_train,
big_X_test,
big_y_test,
ch_num,
class_names=class_names,
addon=''):
features_train = [None] * len(big_X_train)
labels_train = [None] * len(big_y_train)
for i, (X_user, y_user) in enumerate(zip(big_X_train, big_y_train)):
temp = [item for sublist in X_user for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train[i] = temp.reshape(temp.shape[0], 1, ch_num, 1125)
lab = [item for sublist in y_user for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train[i] = np_utils.to_categorical(encoded_Y)
features_test = [None] * len(big_X_test)
labels_test = [None] * len(big_y_test)
for i, (X_user, y_user) in enumerate(zip(big_X_test, big_y_test)):
temp = [item for sublist in X_user for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_test[i] = temp.reshape(temp.shape[0], 1, ch_num, 1125)
lab = [item for sublist in y_user for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_test[i] = np_utils.to_categorical(encoded_Y)
metrics = np.zeros(((len(big_y_train)), 4))
for i in range(len(big_X_train)):
model = EEGNet_org(nb_classes=len(class_names),
Chans=ch_num,
Samples=1125,
dropoutRate=0.2)
model.compile(loss=categorical_crossentropy,
optimizer=Adam(),
metrics=['accuracy'])
filename_ = '{0}{1}{2}'.format(model_name,
'standard_user_{}'.format(i + 1), addon)
metrics[i] = standard_unit(model, features_train[i], labels_train[i],
features_test[i], labels_test[i], filename_,
class_names)
metrics_to_csv(metrics, '{}_Standard_testing_{}'.format(model_name, addon))
def full_freezing(model_name, big_X_train, big_y_train, big_X_test, big_y_test,
class_names=class_names, ch_num=25, dr=0.1, addon=''):
###
### Take all the data for training
###
###
### Train and save the model
###
my_list = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
model = EEGNet_org(nb_classes=2, Chans=ch_num, Samples=1125, dropoutRate=dr)
model.compile(loss=categorical_crossentropy,
optimizer=Adam(), metrics=['accuracy'])
features_train = []
labels_train = []
for i in my_list:
temp = [item for sublist in big_X_train[i] for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train.append(temp.reshape(temp.shape[0], 1, ch_num, 1125))
lab = [item for sublist in big_y_train[i] for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train.append(np_utils.to_categorical(encoded_Y))
# Also add the testing data to increase the size of training data
temp = [item for sublist in big_X_test[i] for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train.append(temp.reshape(temp.shape[0], 1, ch_num, 1125))
lab = [item for sublist in big_y_test[i] for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train.append(np_utils.to_categorical(encoded_Y))
# Flatten the data for training
full_features = np.vstack(features_train)
full_labels = np.vstack(labels_train)
filename_ = '{0}{1}{2}'.format(model_name, '_Full_Freezing_2mvt', addon)
full_freezing_unit(model, full_features, full_labels,
filename_, class_names)
def opt_Dropout_rate_CV_EEGNet(dropout_start, dropout_stop, model_name, big_X_train, big_y_train,
big_X_val, big_y_val, class_names=class_names, ch_num=25):
for i in np.arange(dropout_start, dropout_stop, 0.01):
EEGnet = EEGNet_org(nb_classes=4, Chans=ch_num, Samples=1125, dropoutRate=i)
standard_all(EEGnet, model_name, big_X_train, big_y_train, big_X_val, big_y_val,
ch_num=ch_num, class_names=class_names,
addon='Dropout_{0:.2f}'.format(i))
def splitted_layers(model_name,
big_X_train,
big_y_train,
big_X_test,
big_y_test,
class_names=class_names,
ch_num=25,
dr=0.1,
addon=''):
###
### First create the sequence of training users and single test user
###
list_element = []
my_list = [0, 1, 2, 3, 4, 5, 6, 7, 8]
series = []
for idx in range(90):
if idx % 10 != 0:
list_element.append(my_list[idx % len(my_list)])
elif idx % 10 == 0:
series.append(list_element)
list_element = []
series[0] = list_element
features_train = []
labels_train = []
metrics = np.zeros(((len(my_list)), 4))
###
### Iterate trough the sequences
###
for user_list in series:
print('Starting Splitted learning for user {}'.format(user_list[-1]))
model = EEGNet_org(nb_classes=2,
Chans=ch_num,
Samples=1125,
dropoutRate=dr)
model.compile(loss=categorical_crossentropy,
optimizer=Adam(),
metrics=['accuracy'])
# Create the first training data
for i in user_list[:-1]:
temp = [item for sublist in big_X_train[i] for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train.append(temp.reshape(temp.shape[0], 1, ch_num, 1125))
lab = [item for sublist in big_y_train[i] for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train.append(np_utils.to_categorical(encoded_Y))
# Also add the testing data to increase the size of training data
temp = [item for sublist in big_X_test[i] for item in sublist]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train.append(temp.reshape(temp.shape[0], 1, ch_num, 1125))
lab = [item for sublist in big_y_test[i] for item in sublist]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train.append(np_utils.to_categorical(encoded_Y))
###
### Create the second training data (of the specific user)
###
temp = [
item for sublist in big_X_train[user_list[-1]] for item in sublist
]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_train_2 = temp.reshape(temp.shape[0], 1, ch_num, 1125)
lab = [
item for sublist in big_y_train[user_list[-1]] for item in sublist
]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_train_2 = np_utils.to_categorical(encoded_Y)
# Flatten the data for training
full_features = np.vstack(features_train)
full_labels = np.vstack(labels_train)
###
### Create the testing data (of the specific user)
###
temp = [
item for sublist in big_X_test[user_list[-1]] for item in sublist
]
temp = np.asanyarray(temp)
temp = np.swapaxes(temp, 1, 2)
features_test_2 = temp.reshape(temp.shape[0], 1, ch_num, 1125)
lab = [
item for sublist in big_y_test[user_list[-1]] for item in sublist
]
# encode class values as integers
encoder = LabelEncoder()
encoder.fit(lab)
encoded_Y = encoder.transform(lab)
# convert integers to dummy variables (i.e. one hot encoded)
labels_test_2 = np_utils.to_categorical(encoded_Y)
filename_ = '{0}{1}{2}'.format(model_name,
'Splitted_{}'.format(user_list[-1] + 1),
addon)
metrics[user_list[-1]] = freezing_unit(model, full_features,
full_labels, features_train_2,
features_test_2, labels_train_2,
labels_test_2, filename_,
class_names)
metrics_to_csv(metrics,
'{}_Splitted_Learning_{}'.format(model_name, filename_))
train_list = glob.glob('../Data/4mvt/A*E.mat')
test_list = glob.glob('../Data/4mvt/A*T.mat')
train_list.sort()
test_list.sort()
tested_channels = [1, 3, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 17]
# Freezing Learning
big_X_train, big_y_train = prepare_data_standard_from_list(
load_simplify_data(train_list, True, False, tested_channels))
big_X_test, big_y_test = prepare_data_standard_from_list(
load_simplify_data(test_list, True, False, tested_channels))
EEGnet = EEGNet_org(nb_classes=4, Chans=13, Samples=1125, dropoutRate=0.1)
standard_all(EEGnet,
'EEG_net',
big_X_train,
big_y_train,
big_X_test,
big_y_test,
class_names=['A', 'B', 'C', 'D'],
ch_num=13,
addon='_13_channels')
full_distributed(model_name,
big_X_train,
big_y_train,
big_X_test,