Network Intrusion Detection Method Based on FCWGAN and BiLSTM

Table 9

Comparison of data enhancement algorithms on NSL-KDD dataset.

Algorithm	Evaluation metrics
	Accuracy	Precision	Recall	F1-score	Time (s)
ROS-BiLSTM	89.18 ± 0.35	90.34 ± 0.40	88.61 ± 0.35	89.46 ± 0.37	4
ADASYN-BiLSTM	92.95 ± 0.24	93.12 ± 0.27	92.61 ± 0.21	92.86 ± 0.25	5
SMOTE-BiLSTM	93.66 ± 0.28	94.63 ± 0.34	93.14 ± 0.26	93.88 ± 0.30	3
WGAN-BiLSTM	96.56 ± 0.23	96.71 ± 0.28	95.65 ± 0.21	96.20 ± 0.26	7
Model in this paper	99.57 ± 0.21	99.55 ± 0.20	99.47 ± 0.17	99.51 ± 0.18	2