arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2207.03579cs.SIcs.AI
收藏

DGraph:一个用于图异常检测的大规模金融数据集

DGraph: A Large-Scale Financial Dataset for Graph Anomaly Detection

Zhejiang University(浙江大学);Finvolution Group(信也科技集团);Fudan University(复旦大学);École Polytechnique(巴黎综合理工学院)

Xuanwen Huang, Yang Yang, Yang Wang, Chunping Wang, Zhisheng Zhang, Jiarong Xu, Lei Chen, Michalis Vazirgiannis

AI总结:

针对图异常检测领域数据集多样性不足的问题,本文提出了金融领域大规模真实动态图数据集DGraph,通过观测和实验验证其可推动该领域研究及异常节点深度探索。

摘要:

图异常检测(Graph Anomaly Detection, GAD)因其实用性和理论价值,近来已成为研究热点。由于GAD强调应用场景且异常样本十分稀少,丰富其数据集的多样性是一项基础工作。为此,本文提出了DGraph——一个金融领域的真实动态图数据集。DGraph克服了现有GAD数据集的诸多局限,包含约300万个节点、400万条动态边以及100万个有真实标签的节点。本文对DGraph进行了全面观测,发现异常节点与正常节点通常在结构、邻域分布和时间动态特征上存在差异,同时表明无标签节点对检测欺诈者也至关重要。此外,本文在DGraph上开展了大量实验。观测结果与实验均证明,DGraph能够推动GAD研究的发展,支持对异常节点的深度探索。

英文摘要:

Graph Anomaly Detection (GAD) has recently become a hot research spot due to its practicability and theoretical value. Since GAD emphasizes the application and the rarity of anomalous samples, enriching the varieties of its datasets is fundamental work. Thus, this paper present DGraph, a real-world dynamic graph in the finance domain. DGraph overcomes many limitations of current GAD datasets. It contains about 3M nodes, 4M dynamic edges, and 1M ground-truth nodes. We provide a comprehensive observation of DGraph, revealing that anomalous nodes and normal nodes generally have different structures, neighbor distribution, and temporal dynamics. Moreover, it suggests that unlabeled nodes are also essential for detecting fraudsters. Furthermore, we conduct extensive experiments on DGraph. Observation and experiments demonstrate that DGraph is propulsive to advance GAD research and enable in-depth exploration of anomalous nodes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑