Table 3 The mean evaluation metrics were derived from running the model 100 times to discriminate between normal and abnormal user profiles. This rigorous testing approach ensured the reliability of the performance indicators by accounting for variability

From: A novel deep synthesis-based insider intrusion detection (DS-IID) model for malicious insiders and AI-generated threats

Accuracy

Cohen’s kappa

TPR

FPR

FAR

Recall

Precision

F1 score

AUC

97.31%

0.81

0.86

0.019

0.14

0.86

0.79

0.82

0.99

Back to article page