Table 1 Quantitative results on NSL-KDD.

From: A novel multi-scale network intrusion detection model with transformer

Method

Binary-classification

Five-classification

Accuracy

Precision

Recall

F1-score

Accuracy

Precision

Recall

F1-score

CNN

91.86

90.93

87.82

89.21

85.12

86.62

85.13

85.40

RNN

97.64

97.45

96.39

96.91

93.16

94.02

91.40

92.56

LSTM

98.71

98.32

98.41

98.36

95.51

95.38

94.44

94.85

ViT

98.79

98.63

98.26

98.44

97.80

97.45

97.83

97.62

Liu et al. 49

92.90

89.92

98.57

94.05

85.24

–

–

–

ANN 50

97.50

99.00

96.70

95.70

–

–

–

–

AE 51

90.61

86.83

98.43

92.26

–

–

–

–

Ours

99.25

99.07

99.02

99.05

99.16

99.01

99.17

99.09

  1. The values are expressed in %, and the best one is in bold.