activation=activation,
kernel_initializer=kernel_initializer))
classificador.add(Dropout(dropout))
classificador.add(Dense(units=1, activation='sigmoid'))
opt = keras.optimizers.Adam(learning_rate=learning_rate,
decay=decay,
beta_1=beta_1,
beta_2=beta_2)
classificador.compile(optimizer=opt, loss=loss, metrics=['binary_accuracy'])
classificador.fit(x_train, y_train, batch_size=batch_size, epochs=epochs)
qtd_param = classificador.count_params()
print('Number of Parameters: ', qtd_param)
print('Calculating the ROC curve...')
previsoes_rna = classificador.predict(x_valid)
prob_rna = previsoes_rna
previsoes_rna = (previsoes_rna > 0.5)
previsoes_num_rna = []
for i in previsoes_rna:
if i:
previsoes_num_rna.append(1)
else:
previsoes_num_rna.append(0)
previsoes_rna = np.array(previsoes_num_rna)
