Table 3.

Hyperparameters of model training

Model	Layers	Epochs	Batch size	Optimizer
CNN	16	20	32	Adam
MLP	16	20	32	Adam
Bi-LSTM	16	20	32	Adam
Bi-LSTM-CNN	32	20	32	Adam
BERT (variants) -fine tune	768	10	8	Adam
BERT-pre-training	768	10	8	Adam