Yelp-Fraud

该数据集名为Yelp-Fraud，包含了以每个节点代表一条评论的评审信息。其中，诸如产品和用户ID等属性作为节点属性存储。该图的建模方式与PFFD相似，因此可以应用类似的检测技术。此外，中小企业对节点进行了欺诈标注，即是否为欺诈评论的标签。该数据集还被用于评估人机交互反馈对检测欺诈评论的影响。任务目标是进行评论欺诈检测。

This dataset is named Yelp-Fraud, which contains review information where each node represents a single review. Attributes including product and user IDs are stored as node attributes. The graph is modeled in a manner similar to PFFD, enabling the application of similar detection technologies. Additionally, nodes are annotated with fraud labels by small and medium-sized enterprises (SMEs), which indicate whether the corresponding review is fraudulent. This dataset has also been used to evaluate the impact of human-machine interaction feedback on fraudulent review detection. The task objective of this dataset is review fraud detection.

搜集汇总

背景与挑战

背景概述

Yelp-Fraud数据集包含以节点形式存储的评论信息，具有产品和用户ID等属性，并标注了欺诈评论标签。该数据集适用于欺诈检测技术研究，特别是评估人机交互反馈对检测效果的影响。