Table 8 Comparison of implemented methods for classification.

From: Optimized extreme learning machines with deep learning for high-performance network traffic classification

Evaluation Metrics

RFC

MLP

GA-ELM

PSO-ELM

TP

10,813

10,272

11,007

10,992

FP

344

885

150

165

TN

8128

7913

8151

8391

FN

342

557

319

79

Accuracy (%)

96.504

92.653

97.610

98.756

Sensitivity

0.969

0.948

0.971

0.992

Specificity

0.959

0.899

0.981

0.980

Precision

0.969

0.920

0.986

0.985

Recall

0.969

0.948

0.971

0.992

F1 Score

0.969

0.934

0.979

0.989

Training time (minutes)

2.650

26.133

35.566

27.616

Prediction time (microseconds)

117.604

23.706

13.894

14.740

Prediction accuracy stability

No

No

No

Yes