Explainable Fraud Detection for Few Labeled Time Series Data

Table 2

The experimental results on the PaySim simulation dataset.

		(%)	Recall (%)	Precision (%)	AUC
AP	0.5481	80.22 ± 2.87	59.47	50.82	0.7006
SVM	0.5911	75.68 ± 2.39	48.91	69.17	0.7343
RF	0.6837	82.74 ± 1.39	59.60	80.18	0.8171
HOBA	0.6439	83.95 ± 2.12	63.23	66.67	0.7658
Ours	0.7789	83.63 ± 1.08	59.04	98.72	0.8910