套现交易识别-社区发现算法

基于短时间内银行卡在不同POS机的消费轨迹，建立POS机关系网络并划分社团，结合社团及商户交易特征识别疑似套现团伙，发现更隐蔽的套现行为（如小额高频循环套现、模拟正常消费等）。识别了大量的疑似套现POS机和疑似套现的信用卡，接近市场对套现形势的预期。

Based on the consumption trajectories of bank cards across different POS terminals within a short time frame, a POS terminal relationship network is constructed and community partitioning is performed. Suspected cash-out rings are identified by combining community attributes and merchant transaction characteristics, and more concealed cash-out behaviors such as small-value high-frequency circular cash-out and simulated legitimate consumption are discovered. A large number of suspected cash-out POS terminals and suspected cash-out credit cards have been identified, which aligns with market expectations regarding the cash-out landscape.