Skip to main content

Table 13 Comparison analysis of UNSW-NB15 dataset

From: Machine learning-based network intrusion detection for big and imbalanced data using oversampling, stacking feature embedding and feature extraction

SI. No.

Authors

Data balancing

Dimension reduction

Algorithm

Selected reature

Binary Acc(%)

Multilabel Acc(%)

1

[25]

SMOTE

–

ELM

–

98.43

–

2

[26]

–

XGB

DT, ANN

19

90.85 (DT)

77.51 (ANN)

3

[29]

–

MQTT+TCP

RF

–

98.67

97.37

4

[44]

–

WFEU

FFDNN

22

87.10

77.16

5

[38]

–

–

DNN

–

91.50

–

6

[37]

–

–

ANN

–

–

97.89

7

[39]

STL

–

LSTM+CNN

–

–

–

8

[22]

SGM

–

CNN

–

–

96.54

9

[45]

–

–

CNN-WDLSTM

–

97.17

98.43

10

Our Proposal

RO

SFE-PCA

RF

10

99.59

99.95

11

Our Proposal

RO

SFE-PCA

ET

10

99.59

99.95