Skip to main content

Table 20 Maximum AUC by classifier for the data exfiltration attack type; mean and standard deviations of AUC, (10 iterations of 5-fold cross-validation)

From: IoT information theft prediction using ensemble feature selection

Experiment name Mean SD
AUC AUC
LR Tuned RUS 1:1 7 Agree 0.96968 0.02346
CatBoost Tuned RUS 1:9 4 Agree 0.99076 0.01461
DT RUS 1:3 4 Agree 0.98398 0.01531
Light GBM Tuned RUS 1:3 5 Agree 0.98947 0.01328
RF RUS 1:1 6 Agree 0.99024 0.01120
MLP RUS 1:3 4 Agree 0.96914 0.02814
NB Tuned RUS 1:3 7 Agree 0.97958 0.01523
XGBoost RUS 1:3 6 Agree 0.98737 0.01497