Table 7 Performance metrics of different models.

From: An incremental adversarial training method enables timeliness and rapid new knowledge acquisition

 

Accuracy

Precision

Recall

F1-score

DNN

0.9627

0.9628

0.9645

0.9632

BiLSTM

0.9640

0.9645

0.9654

0.9645

BiLSTM-MultiheadAttention

0.9587

0.9588

0.9613

0.9597

CNN-LSTM

0.9560

0.9553

0.9562

0.9557

CNN-LSTM-ATT

0.9307

0.9289

0.9307

0.9294

NHANet

0.9840

0.9842

0.9837

0.9839

Back to article page