| Project Reference: | ITP/050/22LP |
| Project Title: | Data Analytics on Faster Payment System |
| Hosting Institution: | LSCM R&D Centre (LSCM) |
| Abstract: | Since September 2018, the Hong Kong Monetary Authority (HKMA) has launched the Faster Payment System (FPS) to address the increasing market needs for more efficient retail payment services. The FPS operates on 24x7 basis to enable customers of banks and e-wallet operators to make instant cross-bank/e-wallet transactions in Hong Kong dollar (HKD) and Renminbi (RMB) easily in Hong Kong. Nowadays, FPS has been used extensively in Hong Kong and attributed to more than 263 million electronic transactions in 2021. With the tremendous volume of FPS transaction data, it is a challenge to identify any anomaly transaction that needs to call for further investigation. We propose to use a machine learning approach to tackle the problem of identifying anomaly transaction from FPS data. In this project, an FPS anomaly transaction detection module based on Simplified Graph Convolution Network methodology will be built. The model will detect anomaly based on the pattern of FPS transaction flow among different FPS accounts. Detecting anomaly transaction is important for anti-money laundering purpose but it is a difficult task to capture these transactions from a large volume of transaction data. Our project result should help HKMA or any financial institutions that need to address this issue. |
| Project Coordinator: | Dr Chung-Dak SHUM |
| Approved Funding Amount: | HK$ 2.79 M |
| Project Period: | 1 Feb 2023 - 31 Jul 2024 |
