MS_HGNN: a hybrid online fraud detection model to alleviate graph-based data imbalance. Issue 1 (31st December 2023)

• Record Type: Journal Article
• Authors: Long, Jing; Fang, Fei; Luo, Cuiting; Wei, Yehua; Weng, Tien-Hsiung
• Is Part Of: Connection science. Volume 35:Issue 1(2023)
• Publication Date: 2023-12-31
• Subjects: Fraud detection -- Data imbalance -- Graph neural network -- Online transaction -- Reinforcement learning; Neural computers -- Periodicals; Artificial intelligence -- Periodicals; Cognitive science -- Periodicals; Connectionism -- Periodicals; 006.3
• DOI: 10.1080/09540091.2023.2191893
• Languages: English
• ISSNs: 0954-0091

Abstract

Abstract : Online transaction fraud has become increasingly rampant due to the convenience of mobile payment. Fraud detection is critical to ensure the security of online transactions. With the development of graph neural network, researchers have applied it to the field of fraud detection. The existing fraud detection methods will solve the class imbalance by sampling, but they do not fully consider the various imbalances in the heterogeneous graph, and the data imbalance will directly affect the performance of the model. This work proposes a hybrid graph neural network model for online fraud detection to address this issue. The three types of imbalance in online transactions are feature imbalance, category imbalance, and relation imbalance, and they are all addressed in the proposed model. The entities with the feature most closely related to the fraudsters will be determined for the feature imbalance, and samples will be taken for further identification in the subsequent training phase. The hybrid model then uses under-sampling in combination with the long-distance sampling to find nodes with high similarity of features for the category imbalance. Finally, we propose a reward/punishment mechanism based on reinforcement learning for relation imbalance, which uses the threshold created by training as the sampling weight between relations. This paper conducts experiments on the public datasets Amazon and Yelp. The experimental results show that the model proposed is 5.61% higher than the best model in the comparison model on Amazon dataset, and 1.58% higher on Yelp dataset.