贝宝如何利用实时图形数据库和图形分析来打击欺诈行为

Prevent organized and repeat fraudsters by using a home-grown graph platform

通过使用本土化的图形平台，防止有组织的和重复的欺诈者

By Quinn Zuo, David Zhang, Yan Zhang, and Nitin Sharma

作者：Quinn Zuo,David Zhang,Yan Zhang, andNitin Sharma

Introduction

简介

-commerce has grown exponentially in recent years, fueled by the COVID-19 pandemic. Digital payments are the lifeblood of e-commerce, and with their rise, comes the rise of payments fraud. Detection and prevention of fraud associated with digital payments is complex and ever-evolving, since fraudsters are constantly trying to find new ways to monetize. The payments industry itself is also a targeted domain by organized and repeat fraudsters.

-近年来，在COVID-19大流行病的推动下，电子商务呈指数级增长。数字支付是电子商务的命脉，而随着数字支付的兴起，支付欺诈也随之兴起。侦查和预防与数字支付相关的欺诈行为是复杂和不断发展的，因为欺诈者不断试图寻找新的盈利方式。支付行业本身也是有组织和重复欺诈者的目标领域。

PayPal supports over 400 million active consumers and merchants worldwide. Every minute there are several thousand payment transactions. The relationships in this payment network are strong predictors of behaviors and great risk indicators that can reveal fraud. However, these relationships are too huge to traverse and analyze if they were to be stored in a relational database. Graph database, on the other hand, flattens the view and treats relationships (connections) as a first-class citizen and efficiently traverses through connections. Graph analysis can then be performed on top of the graph data to derive insights, such as if a buyer account is compromised and could lead to account takeover, or if a buyer is likely to interact with a risky seller in the next three months, etc.

PayPal支持全球超过4亿活跃的消费者和商家。每分钟都有几千笔支付交易。这个支付网络中的关系是对行为的有力预测，也是能够揭示欺诈的巨大风险指标。然而，如果将这些关系存储在关系型数据库中，则太过庞大，无法进行穿越和分析。另一方面，图数据库将视图扁平化，并将关系（连接）作为一级公民对待，并有效地穿越连接。然后，可以在图形数据的基础上进行图形分析，以获得洞察力，例如，一个买方账户是否受到损害并可能导致账户被接管，或者一个买方是否有可能在未来三个月内与一个有风险的卖家互动，等等。

Graph Database and Graph Analysis

图形数据库和图形分析

Graph is a universal language for describing relationships between data points (entities) with two simpl...