Search

Graph-based fraud detection methods have recently attracted much attention due to the rich relational information of graph-structured data, which may facilitate the detection of fraudsters. However, the GNN-based algorithms may exhibit unsatisfactory performance faced with graph heterophily as the fraudsters usually disguise themselves by deliberately making extensive connections to normal users. In addition to this, the class imbalance problem also causes GNNs to overfit normal users and perform poorly for fraudsters. To address these problems, we propose an Imbalanced Graph Structure Learning framework for fraud detection (IGSL for short). Specifically, nodes are picked with a devised multi-relational class-balanced sampler for mini-batch training. Then, an iterative graph structure learning module is proposed to iteratively construct a global homophilic adjacency matrix in the embedding domain. Further, an anchor node message passing mechanism is proposed to reduce the computational complexity of the constructing homophily adjacency matrix. Extensive experiments on benchmark datasets show that IGSL achieves significantly better performance even when the graph is heavily heterophilic and imbalanced. [ABSTRACT FROM AUTHOR]

The development of mobile communication technologies has enriched human interaction and inevitably led to the proliferation of telecom fraud. While recent studies have paid attention to collaborative fraud in telecom networks, they have examined only voice networks and neglected the collaborative fraud phenomenon in Short Message Service (SMS) networks and across voice and SMS networks. To address this issue, we investigated a telecom dataset with 5.01 million call records and 6.84 million SMS records between 6,106 users and 2.02 million recipients over eight months. On the one hand, we find differences in the evolution pattern of SMS behaviour between fraudsters and normal users. On the other hand, we also confirm the existence of intra-network collaborative fraud in voice networks and SMS networks and inter-network collaborative fraud across voice and SMS networks. To this end, we propose a novel telecom fraud detector, MLCG-FD, which models dynamic individual behaviours and multi-network collaborative relationships. First, a multi-network latent collaborative network (LGC) extractor is designed to extract two intra-network LCGs and one inter-network LCG from voice and SMS networks. Then, we design two behavioural encoders to encode individual voice and SMS behaviour patterns. Finally, a network-based hybrid-pass filter aggregates higher-order neighbour features from different networks. Experiments conducted on two real telecom datasets show that MLCG-FD significantly improves over state-of-the-art methods. • Telecom fraudsters commonly employ multi-network collaborative fraud strategies. • Individual SMS behaviour between fraudsters and normal users are different. • Latent collaborative graphs have a high heterophily problem. • Network-based hybrid-pass filter obtain the optimal combination of frequencies. [ABSTRACT FROM AUTHOR]

The development of telecom technology not only facilitates social interactions but also inevitably provides the breeding ground for telecom fraud crimes. However, telecom fraud detection is a challenging task as fraudsters tend to commit co-fraud and disguise themselves within the mass of benign ones. Previous approaches work by unearthing differences in calling sequential patterns between independent fraudsters, but they may ignore synergic fraud patterns and oversimplify fraudulent behaviors. Fortunately, graph-like data formed by traceable telecom interaction provides opportunities for graph neural network (GNN)-based telecom fraud detection methods. Therefore, we develop a latent synergy graph (LSG) learning-based telecom fraud detector, named LSG-FD, to model both sequential and interactive fraudulent behaviors. Specifically, LSG-FD introduces (1) a multi-view LSG extractor to reconstruct synergy relationship-oriented graphs from the meta-interaction graph based on second-order proximity assumption; (2) an LSTM-based calling behavior encoder to capture the sequential patterns from the perspective of local individuals; (3) a dual-channel based graph learning module to alleviate the disassortativity issue (caused by the camouflages of fraudsters) by incorporating the dual-channel frequency filters and the learnable controller to adaptively aggregate high- and low-frequency information from their neighbors; (4) an imbalance-resistant model trainer to remedy the graph imbalance issue by developing a label-aware sampler. Experiment results on the telecom fraud dataset and another two widely used fraud datasets have verified the effectiveness of our model. [ABSTRACT FROM AUTHOR]