Facilitating User Authorization from Imbalanced Data Logs of Credit Cards Using Artificial Intelligence

Table 6

Sensitivity, specificity, precision, and F1 scores obtained on Dataset B executing RUSBoost, customized RUSBoost, DT, LR, MLP, KNN, RF, AdaBoost, and SVM.

Model name	Sensitivity	Specificity	Precision	F1 score
RUSBoost	34.6	98.3	85.9	59.4
Customized RUSBoost	99.6	98.7	95.7	97.6
DT	40.6	81.0	49.5	50.7
LR	23.6	97.0	69.6	35.0
MLP	38.5	93.2	61.4	47.3
KNN	37.8	89.4	50.0	43.1
RF	5.5	99.2	68.2	10.2
AdaBoost	30.8	95.8	67.3	42.3
SVM	33.2	95.2	67.8	44.5