Skip to main content
. 2023 Oct 30;9(11):e21377. doi: 10.1016/j.heliyon.2023.e21377

Table 2.

AUC of the proposed HHOSSA Hybrid Classifier and other existing methods based on PRC.


Methods
Attacks
Benign traffic
Data exfiltration
Established foothold
Lateral movements
Reconnaissance
AUC %
SVM 64.081 63.751 64.960 59.925 62.276
SAE-LSTM 65.293 64.434 66.715 60.560 62.939
KNN 67.003 66.489 68.097 62.471 64.939
Hybridized CNN and Bi-LSTM 68.270 68.863 68.269 64.679 67.249
Light GBM 69.384 70.044 69.328 65.777 68.398
Adaboost 71.159 71.374 71.564 67.014 69.692
Bi-LSTM 72.905 72.873 73.291 68.409 71.151
PSO-BiLSTM 75.536 73.746 77.588 69.220 72.000
SSO-BiLSTM 76.558 75.342 77.714 70.704 73.553
HHO-BiLSTM 78.554 79.301 78.501 74.386 77.405
HHOSSA-Hybrid Classifier 80.737 83.094 79.097 77.914 81.096