Skip to main content

Table 8 Performance of the proposed model under different thresholds

From: Network abnormal traffic detection method based on fusion of chord similarity and multiple loss encoder

AE threshold

Accuracy (%)

Precision (%)

Recall (%)

F1 score (%)

Min loss

83.904

78.604

98.598

87.466

Max loss

91.644

91.714

93.812

92.744

Mean loss

87.224

83.036

97.64

89.722

  1. The best-performing AE threshold is marked with bold styles