model.fit(train_data, train_labels, batch_size=batchsize, epochs=10, verbose=2)
test_predict = model.predict(test_data, batch_size=batchsize, verbose=0)
fpr, tpr, thresholds = roc_curve(flatten_test_labels, test_predict[:, 1], pos_label=1)
test_auc_value = auc(fpr, tpr)
auc_values.append(test_auc_value)
plt.plot(range(0,500,10),auc_values)
plt.title('Test AUC for filtered EEG')
plt.ylabel('AUC')
plt.xlabel('epoch')
plt.show()
"""
# model.save('***.h5')
# generate prediction probabilities
train_predict = model.predict(train_data, batch_size=batchsize, verbose=0)
test_predict = model.predict(test_data, batch_size=batchsize, verbose=0)
print(model.evaluate(test_data, test_labels, batch_size=batchsize))
# Calculate AUC
fpr, tpr, thresholds = roc_curve(flatten_train_labels,
train_predict[:, 1],
pos_label=1)
print("Training auc:")
train_auc_value = auc(fpr, tpr)
print(train_auc_value)
fpr, tpr, thresholds = roc_curve(flatten_test_labels,
test_predict[:, 1],
pos_label=1)
print("Testing auc:")
test_auc_value = auc(fpr, tpr)
print(test_auc_value)
validation_data=(X_train_valid[test],
y_train_valid[test]),
callbacks=[checkpointer],
class_weight=class_weights)
#
print(
'------------------------------------------------------------------------'
)
print(f'Training for fold {fold_no} ...')
# Evaluation de la performance sur l'ensemble test
probs = model.predict(X_train_valid[test])
preds = probs.argmax(axis=-1)
scores = model.evaluate(X_train_valid[test],
y_train_valid[test],
verbose=0)
auc = roc_auc_score(y_train_valid[test], preds)
acc_per_fold.append(scores[1] * 100)
auc_per_fold.append(auc)
# On passe à un autre pli
fold_no = fold_no + 1
# Evaluation finale sur l'ensemble de test
probs = model.predict(X_test)
preds = probs.argmax(axis=-1)
acc = np.mean(preds == y_test.argmax(axis=-1))
optimizer='adam',
metrics=['accuracy'])
history = model.fit(train_data,
train_labels,
batch_size=batchsize,
epochs=300,
verbose=2)
# model.save('***.h5')
# generate prediction probabilities
train_predict = model.predict(train_data, batch_size=batchsize, verbose=0)
validation_predict = model.predict(validation_data,
batch_size=batchsize,
verbose=0)
test_predict = model.predict(test_data, batch_size=batchsize, verbose=0)
validation_accuracy = model.evaluate(validation_data,
validation_labels,
batch_size=batchsize)
print(validation_accuracy)
test_accuracy = model.evaluate(test_data,
test_labels,
batch_size=batchsize)
print(test_accuracy)
# Calculate AUC
fpr, tpr, thresholds = roc_curve(flatten_train_labels,
train_predict[:, 1],
pos_label=1)
print("Training auc:")
train_auc_value = auc(fpr, tpr)
print(train_auc_value)