发表机构
Emory University; University of Illinois at Chicago（埃默里大学; 伊利诺伊大学芝加哥分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对时序图边级欺诈检测中现有共形预测器预测集低效的问题，提出ProtoCP框架，通过原型和邻域相对得分机制优化校准，在四个基准上实现目标覆盖率且预测集更小。
AI 中文摘要
共形预测（Conformal Prediction，CP）提供无分布的覆盖率保证，已成为不确定性量化的原则性工具。在时序交互图的边级欺诈检测中，误报和漏报均带来巨大成本，此类覆盖率保证对风险感知决策尤为有吸引力。然而，直接应用现有图共形预测器会产生低效的预测集，原因在于欺诈数据的两个常见特性：欺诈交互常嵌入良性主导的邻域中，稀释了校准信号；极端类别不平衡导致校准划分中标记欺诈的支持样本稀少，进而产生过于保守的类别条件阈值。为解决这些问题，我们提出ProtoCP，一种用于时序图边级欺诈检测的共形预测框架。ProtoCP通过将校准聚焦于欺诈相关子图上下文，在类别不平衡和时序漂移下生成更稳定的非一致性得分，从而提升校准效率。具体而言，它利用学习到的原型抑制校准上下文中良性主导的噪声，并引入带时序得分扩散的邻域相对得分机制，实现稳定的类别条件校准。在四个欺诈基准（YelpChi、S-FFSD、FTFD和BankSim）上的实验表明，ProtoCP在达到目标覆盖率的同时，生成的预测集始终比最先进的基线更小。我们的代码可在此URL获取。
英文摘要
Conformal prediction (CP) provides distribution-free coverage guarantees and has emerged as a principled tool for uncertainty quantification. In edge-level fraud detection on temporal interaction graphs, where false positives and false negatives both carry substantial cost, such coverage guarantees are particularly appealing for risk-aware decision making. However, directly applying existing graph conformal predictors yields inefficient prediction sets due to two recurring properties of fraud data. Fraudulent interactions are often embedded in benign-dominated neighborhoods that dilute calibration signals, while extreme class imbalance leaves scarce labeled-fraud support in the calibration split and leads to overly conservative class-conditional thresholds. To address these issues, we propose ProtoCP, a conformal prediction framework for edge-level fraud detection on temporal graphs. ProtoCP improves calibration efficiency by focusing calibration on fraud-relevant subgraph context and producing more stable nonconformity scores under class imbalance and temporal drift. Specifically, it leverages learned prototypes to suppress benign-dominated noise in the calibration context and introduces a neighborhood-relative scoring mechanism with temporal score diffusion for stable class-conditional calibration. Experiments on four fraud benchmarks (YelpChi, S-FFSD, FTFD, and BankSim) show that ProtoCP achieves the target coverage with consistently smaller prediction sets than state-of-the-art baselines. Our codes are available at https://github.com/Picard1701ent/ProtoCP.git
CommentsAccpeted by KDD 2026
