大规模风险商品检测数据集

本数据集来自“ICDM 2022 : 大规模电商图上的风险商品检测挑战赛”初赛数据：https://tianchi.aliyun.com/competition/entrance/531976/information ， 由阿里安全部门提供。<br />
在电商平台上，商品是最主要的内容之一。风险商品检测旨在识别平台上存在的假货商品、违禁商品等，对维护平台内容信息健康、保护消费者权益起着至关重要的作用。<br />

风险商品检测和其他风控领域一样，面临风险的对抗和变异，如对商品风险内容的刻意隐藏等，而使用平台广泛存在的各类图关系数据可以提供更多证据，提升黑灰产的攻击成本。研究人员可利用本次开放的来源于真实场景中的大规模的、异构的、存在噪声和缺失的图数据，以及极不均衡的样本，进行风险商品检测。<br />

This dataset is derived from the preliminary round data of the "ICDM 2022: Risk Product Detection Challenge on Large-scale E-commerce Graphs", and is provided by the Alibaba Security Department. The official competition information page is: https://tianchi.aliyun.com/competition/entrance/531976/information

On e-commerce platforms, products constitute one of the core content types. Risk product detection aims to identify counterfeit, prohibited and other risky products on the platform, which plays a vital role in safeguarding the health of platform content and information, as well as protecting consumers' legitimate rights and interests.

Similar to other risk control fields, risk product detection faces adversarial risks and variations, such as deliberate obfuscation or hiding of risky product content. Utilizing various graph relational data widely existing on the platform can provide additional evidentiary support and increase the attack cost of black and gray cybercrime organizations. Researchers can employ the open, large-scale, heterogeneous, noisy and incomplete graph data derived from real-world scenarios, as well as highly imbalanced sample sets, to conduct risk product detection studies.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于电商平台风险商品检测的大规模异构图表数据集，包含噪声和缺失数据，样本极不均衡。数据集由阿里安全部门提供，旨在通过图算法识别假货和违禁商品，遵循CC BY-NC 4.0协议。