Page 71 - 《软件学报》2026年第4期

1512 软件学报 2026 年第 37 卷第 4 期

environments, cross-agency data sharing and collaboration are difficult due to the involvement of sensitive user information, which in turn
limits the training effectiveness and generalization performance of the model. Federated learning, as an emerging privacy-preserving
distributed learning paradigm, provides a feasible way for cross-agency collaborative training, but existing graph federated learning
methods are mostly designed for general graph tasks, making them difficult to adapt to the class imbalance and data heterogeneity
problems prevalent in fraud detection, resulting in poor performance in fraud sample identification. To address the above challenges, this
study proposes a risk perception dynamic aggregation graph federated learning method (FedRPDA) for fraud detection, aiming to
effectively deal with complex fraud risk event recognition across organizations. FedRPDA includes two key strategies: the typical risk
dynamic aggregation strategy measures the structural risk intensity of fraudulent nodes in the client graph and combines it with a dynamic
weight mapping mechanism with temporal decay characteristics to adaptively adjust the aggregation weights of clients, thus enhancing the
global model’s ability to discriminate between normal samples and typical fraud samples under heterogeneous data conditions; the
diversified risk average aggregation strategy integrates a variance perturbation-based feature enhancement mechanism for fraud samples
with a global prototype-guided contrastive learning mechanism, which effectively improves the model’s ability to represent structurally
diverse and scarce a typical fraud samples, promotes their convergence toward common anomalies in the feature space, and further
enhances the model’s robustness in recognizing complex fraud risk scenarios. Experimental results on several real-world fraud detection
datasets show that FedRPDA significantly outperforms existing graph federated learning baseline methods in terms of detection
performance and training convergence efficiency, and demonstrates good generalization ability and practical application potential.
Key words: fraud detection; federated learning; graph neural network (GNN); diverse risk

数字信息产业的高速发展在推动社会智能化进程的同时, 也为不法分子提供了前所未有的作案手段. 尤其是
伴随着互联网和人工智能技术的革新, 欺诈行为在金融交易 [1,2] 、社交网络 [3,4] 和评论系统 [5,6] 等多个领域呈现出专
业化、规模化和复杂化的特点. 在金融交易场景中, 信用卡盗刷、洗钱及虚假交易等欺诈行为层出不穷, 其攻击手
段隐蔽、行为迅速, 给银行和支付平台带来了严峻挑战, 甚至威胁金融体系的稳定. 社交网络中的欺诈者依托虚假
身份散播误导性信息, 诱导用户上当受骗, 而虚假内容的辨别难度较大, 导致经济损失和社会风险加剧. 评论系统
中的欺诈行为通过虚假评论、操控评分等手段误导消费者决策, 干扰市场公平竞争, 甚至形成大规模商业操纵. 这
些欺诈模式普遍呈现出隐蔽性强、手段多样、规模化作案等特点, 给传统检测方法 [7,8] 带来了极大挑战. 为此, 学
术界与工业界正积极探索新型检测机制, 以提升欺诈识别的鲁棒性和实时性, 为社会治理与风险防控提供坚实的
技术保障.
近年来, 深度学习技术 [9] 凭借其强大的表征学习能力, 在高维非线性数据建模方面表现出色, 已被广泛应用于
欺诈检测场景. 然而, 欺诈行为往往不仅依赖个体特征, 还涉及多个实体之间的复杂交互, 使得数据天然呈现出图
结构属性, 因此图神经网络 (graph neural network, GNN) [10] 被引入为欺诈检测提供了一种更具表现力的建模范式,
通过多层次信息聚合机制挖掘节点间潜在关联, 有效提升欺诈检测的精度和鲁棒性. 尽管基于 GNN 的欺诈检测
方法取得了显著成效, 但该领域依然面临诸多现实挑战, 其中数据的安全性与隐私保护问题尤为突出. 欺诈检测通
常涉及高度敏感的用户数据, 而受限于隐私保护法规及机构间的数据壁垒, 传统的集中式模型难以直接访问跨平
台的全量数据, 导致模型泛化能力受限. 联邦学习 [11] 作为一种分布式机器学习范式, 能够在保护数据隐私的前提
下实现跨机构协同训练, 有效解决数据孤岛问题. 为此, 研究者将联邦学习与图神经网络相结合, 提出图联邦学习 [12] ,
使参与方能够基于本地的图数据进行模型训练, 通过联邦聚合机制实现模型参数共享与更新, 提升全局检测能力.
尽管图联邦学习在跨机构欺诈检测中展现出巨大潜力, 其在实际应用中仍面临两个关键挑战.
(1) 跨机构数据的高度异构性: 不同机构的图数据在类别分布、节点特征分布及拓扑结构等方面存在显著差
异, 这种数据异构性不仅加剧了跨平台模型的训练难度, 也导致模型难以统一建模并在各场景中保持稳定的泛化
性能.
(2) 欺诈数据的极端类别不平衡性: 欺诈检测任务中的正常活动数量远超异常活动, 导致训练过程中模型倾向
于过度拟合多数类正常样本, 忽视少数类欺诈样本的特征表达, 这种偏向性使得模型难以有效捕捉异常模式, 在实
际检测中引发较高的误报率.
数据异构性是导致联邦学习收敛速度下降和模型性能不佳的主要原因 [13] . 非独立同分布 (Non-IID) 数据会引
发局部模型参数更新的方向不一致, 导致在全局模型聚合过程中产生次优解, 当前主流优化方案主要聚焦于客户