RGH-GNN : Risk-Gated Heterogeneous Graph Neural Network for Financial Fraud Detection

Mythreyee Hari; Subin Sahayam; Vegesna S M Srinivasavarma

cs.LGOct 19, 2026

RGH-GNN : Risk-Gated Heterogeneous Graph Neural Network for Financial Fraud Detection

Authors: Mythreyee Hari, Subin Sahayam, Vegesna S M Srinivasavarma

Organizations: Department of Computer Science and Engineering, Shiv Nadar University, Chennai - 603110, Tamil Nadu, India

Abstract

Credit card fraud is a significant problem for financial institutions, resulting in substantial monetary loss each year. Fraud patterns and methods are continuously evolving, with fraudulent transactions often disguised to resemble legitimate activity. Real world transaction data for fraud detection is severely affected by class imbalance making it harder to detect these transactions. Fraudulent transactions are rarely isolated. Accounts, devices, and other common attributes are often shared across fraudulent activity, revealing relational patterns that individual transaction features alone cannot capture. Graph-based approaches model these relationships explicitly. However, standard graph neural networks aggregate neighbourhood information uniformly, which can dilute the already scarce signal carried by fraudulent transactions and limit detection of camouflaged fraud. This work presents RGH-GNN, a heterogeneous graph neural network that addresses this limitation by incorporating transaction-level risk scores, generated by a supervised LightGBM classifier as edge weights. It is used to guide message passing through a learnable risk-gated aggregation mechanism. It ensures that high-risk transactions contribute more strongly to learned representations rather than being diluted by uniform neighbourhood aggregation. RGH-GNN is evaluated on the IEEE-CIS fraud detection dataset against standard graph neural network models that are trained on the same heterogeneous graph. RGH-GNN outperforms all baselines across every evaluation metric, with an AUC-ROC score of 0.914. This demonstrates that incorporating engineered risk signals into graph-based message passing substantially improves fraud detection performance.

Figures & tables

Figure 1: Methodology Diagram RGH-GNN Pipeline

Figure 2: A subgraph sampled over 200 transactions. Account nodes (blue) are connected to clean transactions (green) and fraudulent transactions (red) using account-transaction edges. Each edge depicts a transaction made using the account it is connected to.

Figure 3: Distribution of LightGBM risk scores by class across all 590,540 transactions. Non-fraud transactions (blue) concentrate near zero (mean score 0.131) while fraud transactions (orange) accumulate near 1.0 (mean score 0.793), a margin of 0.662 between classes, with the optimal classification threshold $\tau=0.75$ separating the two populations.

Figure 4: Forward-pass architecture of the proposed RiskCONV layer

$\tau$	$g_{a}$	F1	$\tau$	$g_{a}$	F1	$\tau$	$g_{a}$	F1
0.1	1.0	0.6318	0.2	20.0	0.6355	0.4	10.0	0.6315
0.1	5.0	0.6360	0.3	1.0	0.6331	0.4	15.0	0.6376
0.1	10.0	0.6352	0.3	5.0	0.6340	0.4	20.0	0.6336
0.1	15.0	0.6359	0.3	10.0	0.6359	0.5	1.0	0.6313
0.1	20.0	0.6303	0.3	15.0	0.6373	0.5	5.0	0.6340
0.2	1.0	0.6378	0.3	20.0	0.6358	0.5	10.0	0.6315

Table 1: Grid search results for

\tau

and

g_{a}

, reporting validation F1 score. The best configuration (

\tau=0.4

g_{a}=5.0

) is shown in bold.

$w^{+}$	1	1.5	2	3	5
F1	0.6286	0.6311	0.6203	0.6250	0.6160

Table 2: Validation F1 score across the pos_weight (

w^{+}

) sweep. The best configuration (

w^{+}=1.5

) is shown in bold.

Model	AUC-ROC	AUC-PR	F1	Precision	Recall
RGH-GNN (Full)	$\mathbf{0.914\pm 0.001}$	$\mathbf{0.631\pm 0.002}$	$\mathbf{0.619\pm 0.002}$	$\mathbf{0.734\pm 0.014}$	$\mathbf{0.536\pm 0.009}$
No Gate Aggregation	$0.912\pm 0.001$	$0.624\pm 0.003$	$0.616\pm 0.003$	$0.726\pm 0.016$	$0.535\pm 0.009$
No Risk Scores	$0.885\pm 0.003$	$0.399\pm 0.017$	$0.414\pm 0.010$	$0.400\pm 0.029$	$0.432\pm 0.023$
No Heterogeneity	$0.885\pm 0.002$	$0.407\pm 0.013$	$0.420\pm 0.006$	$0.396\pm 0.010$	$0.448\pm 0.023$
Weighted GraphSAGE	$0.879\pm 0.003$	$0.401\pm 0.019$	$0.418\pm 0.015$	$0.451\pm 0.025$	$0.392\pm 0.035$
Unweighted GraphSAGE	$0.884\pm 0.001$	$0.387\pm 0.011$	$0.405\pm 0.008$	$0.375\pm 0.036$	$0.445\pm 0.035$

Table 3: Ablation study results on the IEEE-CIS fraud detection test set.

Model	Risk Scores	Heterogeneity	Gated Aggregation
RGH-GNN (Full)	✓	✓	✓
No Gated Aggregation	✓	✓	✗
No Risk Scores	✗	✓	✓
No Heterogeneity	✓	✗	✓
Weighted GraphSAGE	✓	✗	✗
Unweighted GraphSAGE	✗	✗	✗

Table 4: Component presence across ablation configurations.

Model	AUC-ROC	AUC-PR	F1	Precision	Recall
Unweighted GraphSAGE [ 15 ]	$0.884\pm 0.001$	$0.387\pm 0.011$	$0.405\pm 0.008$	$0.375\pm 0.036$	$0.445\pm 0.035$
Weighted GraphSAGE [ 15 ]	$0.879\pm 0.003$	$0.401\pm 0.019$	$0.418\pm 0.015$	$0.451\pm 0.025$	$0.392\pm 0.035$
Unweighted GIN [ 39 ]	$0.734\pm 0.002$	$0.146\pm 0.009$	$0.242\pm 0.011$	$0.182\pm 0.017$	$0.368\pm 0.026$
Weighted GIN [ 39 ]	$0.642\pm 0.031$	$0.073\pm 0.004$	$0.186\pm 0.008$	$0.130\pm 0.010$	$0.332\pm 0.021$
Unweighted GAT [ 34 ]	$0.862\pm 0.003$	$0.319\pm 0.009$	$0.363\pm 0.004$	$0.363\pm 0.036$	$0.367\pm 0.030$
Weighted GAT [ 34 ]	$0.865\pm 0.004$	$0.331\pm 0.020$	$0.372\pm 0.016$	$0.371\pm 0.024$	$0.374\pm 0.015$

Table 5: Performance comparison of baseline models against RGH-GNN on the IEEE-CIS fraud detection test set. The best performance is highlighted in bold and the second-best is underlined.

Figure 5: GNN output probability distribution for fraud transactions on the held-out test set. The dashed line marks the classification threshold at 0.94. The majority of fraud transactions accumulate near 1.0 and are correctly classified, while a cluster near 0 represents hard negatives (transactions the model confidently assigned low fraud probability despite being genuinely fraudulent.)

Financial Fraud Detection

Financial fraud detection identifies suspicious transactions, accounts, claims, and behavior to prevent losses and guide investigation.

See all 135 papers

May 12, 2026

Graph-Based Financial Fraud Detection with Calibrated Risk Scoring and Structural Regularization

Financial transaction fraud prevention faces challenges such as complex relationship structures, concealed behavioral patterns, and dynamically changing data distribution. Discrimination models relying solely on independent sample features are insufficient to fully characterize the risks of group collaboration and chain transfers within transaction networks. This paper proposes a graph neural network representation learning and risk discrimination framework for financial transaction fraud prevention. It integrates transaction records and identity information into node attributes and constructs a transaction graph based on shared attributes and interaction consistency to explicitly model inter-transaction relationships. In model design, a multi-layer message passing mechanism is employed to aggregate neighborhood information, learn node embedding representations containing structural context semantics, and output transaction-level fraud probability and risk scores through a lightweight risk discrimination head. A weighted supervision objective is introduced to mitigate training bias caused by class imbalance, and structural consistency regularization constraints are combined to suppress the impact of noisy edges on representation drift, thereby improving the stability and usability of risk characterization. Experiments are conducted on a publicly available financial transaction dataset, comparing various methods in the same direction and comprehensively evaluating them under a unified evaluation protocol. The results show that the proposed method outperforms other methods in risk ranking and probability calibration quality, validating the effectiveness of graph structure modeling and representation learning collaboration in financial transaction fraud prevention.

Yunfei Nie, Jiawei Wang, Ruobing Yan +3

Brandeis University, Waltham, USA · University of California, Los Angeles, USA · Georgetown University, Washington, D.C., USA +2

Graph Neural Networks Graph Representation Learning Financial Fraud Detection

Ask PDF

Sep 7, 2026

Temporal Heterogeneous Graph Transformer for Credit Card Fraud Detection

Credit card fraud detection typically relies on tabular features, while repeated attributes can also provide useful relational signals. This paper proposes THGT-FD, a Temporal Heterogeneous Graph Transformer for Fraud Detection. Each transaction is represented using one transaction token and six types of relation tokens and incorporates Time2Vec encoding into the transaction representation. A Transformer learns the interactions among these tokens within each individual transaction and then outputs a fraud probability. Experiments were conducted on 150,000 transactions sampled from the IEEE-CIS Fraud Detection dataset and chronologically partitioned according to TransactionDT. On the test set, THGT-FD achieved an AUC-ROC of 0.8536, an average precision of 0.4164, and a Recall@5% of 0.4708. The class-weighted histogram-based gradient-boosting baseline achieved an AUC-ROC of 0.8722. The results indicate that relation tokens provide useful information for fraud-risk ranking, although the current model does not yet incorporate entity-level historical aggregation.

Qinwen Yan

University of California, Los Angeles Los Angeles, California, USA

Financial Fraud Detection Transformer Feature Interaction Modeling

Ask PDF

Apr 14, 2026

Graph-Based Fraud Detection with Dual-Path Graph Filtering

Fraud detection on graph data can be viewed as a demanding task that requires distinguishing between different types of nodes. Because graph neural networks (GNNs) are naturally suited for processing information encoded in graph form through their message-passing operations, methods based on GNN models have increasingly attracted attention in the fraud detection domain. However, fraud graphs inherently exhibit relation camouflage, high heterophily, and class imbalance, causing most GNNs to underperform in fraud detection tasks. To address these challenges, this paper proposes a Graph-Based Fraud Detection Model with Dual-Path Graph Filtering (DPF-GFD). DPF-GFD first applies a beta wavelet-based operator to the original graph to capture key structural patterns. It then constructs a similarity graph from distance-based node representations and applies an improved low-pass filter. The embeddings from the original and similarity graphs are fused through supervised representation learning to obtain node features, which are finally used by an ensemble tree model to assess the fraud risk of unlabeled nodes. Unlike existing single-graph smoothing approaches, DPF-GFD introduces a frequency-complementary dual-path filtering paradigm tailored for fraud detection, explicitly decoupling structural anomaly modeling and feature similarity modeling. This design enables more discriminative and stable node representations in highly heterophilous and imbalanced fraud graphs. Comprehensive experiments on four real-world financial fraud detection datasets demonstrate the effectiveness of our proposed method.

Wei He, Wensheng Gan, Philip S. Yu

School of Intelligent Systems Science and Engineering, Jinan University, Zhuhai 519070, China · Department of Computer Science, University of Illinois Chicago, Chicago 60607, USA

Financial Fraud Detection Graph Representation Learning

Ask PDF

Abstract

Financial Fraud Detection

Explore similar work

Graph-Based Financial Fraud Detection with Calibrated Risk Scoring and Structural Regularization

Temporal Heterogeneous Graph Transformer for Credit Card Fraud Detection

Graph-Based Fraud Detection with Dual-Path Graph Filtering