Table 5 Performance analysis for dataset 2.

From: A secure and efficient deep learning-based intrusion detection framework for the internet of vehicles

Metrics

CAN dataset

Proposed

AlexNet

DenseNet

SqueezeNet

GoogleNet

Sensitivity

0.973

0.936

0.960

0.944

0.944

Specificity

0.981

0.944

0.964

0.942

0.939

Accuracy

0.984

0.956

0.970

0.948

0.946

Precision

0.973

0.937

0.967

0.940

0.940

Recall

0.973

0.952

0.962

0.927

0.922

F-measure

0.972

0.905

0.964

0.933

0.930

NPV

0.982

0.945

0.967

0.939

0.939

FPR

0.018

0.035

0.035

0.032

0.032

FNR

0.092

0.152

0.144

0.143

0.142

MCC

0.972

0.931

0.962

0.929

0.929

Training Time (s)

310

380

540

330

420

Testing Time (s)

75

90

100

85

95

Inference Time (ms)

12

18

22

15

20

Back to article page