Fraud Detection on PAYSIM benchmark leaderboard · SOTA2 Research

Fraud Detection on PAYSIM

99F1 ScoreQT-GNN

Evaluation Results

Method	Links
QT-GNN2025.12		99	0.987	98	99.5	98.2	0.04	0.998	0.993
RXT-J2025.12		98.7	0.984	97.7	99.3	97.9	0.046	0.997	0.991
ResNet2025.12		94.2	0.939	91.2	92.3	92.4	0.123	0.98	0.941
TabM2025.12		92	0.92	100	85	—	—	—	—
CapsNet2025.12		92	0.847	87	91.5	91.2	0.2	0.973	0.92
Transformer2025.12		91.6	0.846	87	91.8	91.2	0.199	0.972	0.918
DenseNet1212025.12		89.8	0.825	84.9	89.4	89.1	0.256	0.962	0.896
SVM-GS2025.12		86.7	0.794	81.7	86.3	85.9	0.214	0.943	0.907
Random Forest2025.12		79	0.81	65	100	—	—	—	—
FinFRE-RAG2025.12		73	0.74	65	85	—	—	—	—
FinFRE-RAG2025.12		71	0.72	61	85	—	—	—	—
XGBoost2025.12		68	0.71	55	92	—	—	—	—
SVM2025.12		66.3	0.312	50.9	92.8	56.2	0.981	0.785	0.709
LG2025.12		66.3	0.313	50.9	92.9	56.2	0.977	0.788	0.712
ADA2025.12		66.3	0.313	50.9	92.9	56.2	0.978	0.787	0.711
XGB2025.12		66	0.305	50.6	93	14.2	1.042	0.769	0.697
DTC2025.12		66	0.305	50.6	93	55.7	1.042	0.769	0.697
NB2025.12		66	0.305	50.6	92.9	55.7	1.042	0.769	0.697
LM2025.12		56.2	0.312	56.2	56.2	13.6	1.609	0.632	0.573
FinFRE-RAG2025.12		25	0.36	15	85	—	—	—	—
FinFRE-RAG2025.12		17	0.28	10	85	—	—	—	—
FinFRE-RAG2025.12		11	0.22	6	85	—	—	—	—
FinFRE-RAG2025.12		9	0.19	5	85	—	—	—	—
Qwen3-14B2025.12		0	-0.05	0	46	—	—	—	—
Qwen3-Next-80B2025.12		0	0	0	100	—	—	—	—
Gemma 3-12B2025.12		0	0.01	0	100	—	—	—	—
Gemma 3-27B2025.12		0	0.01	0	100	—	—	—	—
GPT-OSS-20B2025.12		0	-0.03	0	61	—	—	—	—
GPT-OSS-120B2025.12		0	0.02	0	100	—	—	—	—