Google Local Business Reviews and Metadata

该数据集包含了大量关于本地企业的文本评论和五星级评分，这些评论和评分来自数百万用户，同时还包括了用户的人口统计特征。数据集还包括了如用户人口统计信息、GPS坐标、教育和工作历史、以及评论时间戳等特征，并已通过工程化特征进行了扩展，同时为了对比学习还经过了采样处理。规模上，该数据集涵盖了来自450万用户的约1150万条文本评论，这些评论针对的是跨越48,000个类别的310万家本地企业。该数据集的任务是用于协作过滤推荐系统中的虚假文本评论检测。

This dataset contains a large corpus of text reviews and 5-star ratings for local businesses, sourced from millions of users, alongside users' demographic attributes. It further includes additional features such as user demographic information, GPS coordinates, education and employment history, and review timestamps, which have been augmented via engineered features and processed with sampling for contrastive learning applications. In terms of scale, the dataset encompasses approximately 11.5 million text reviews from 4.5 million users, targeting 3.1 million local businesses across 48,000 categories. The target task of this dataset is fake text review detection in collaborative filtering recommendation systems.