Table 3.

Training parameters.

Parameters	Bi-LSTM-Attention	Parameters	1D-CNN
LSTM hidden size	64	Conv num layers	4
LSTM num layers	2	(in, out, kernel size, stride, padding) of layer1	(1,16,3,1,1)
LSTM dropout	0.1	(in, …, padding) of layer2	(16,32,3,1,1)
hidden linear size	256	(in, …, padding) of layer3	(32,32,3,1,1)
linear dropout	0.3	(in, …, padding) of layer4	(32,32,2,1,1)
batch size	20	batch size	32
training epochs	100	training epochs	100