TwiBot-20
收藏资源简介:
TwiBot-20是由西安交通大学创建的大规模Twitter机器人检测基准数据集,包含229,573个用户,33,488,192条推文,8,723,736个用户属性项和455,958个关注关系。该数据集通过广度优先搜索从不同主题的大量种子用户开始,遍历关注关系,确保用户在地理位置和兴趣领域上的多样性,更好地代表当前的Twitter环境。TwiBot-20包括三种用户信息模式,支持单用户二元分类和社区感知方法。数据集的应用领域主要集中在Twitter机器人检测,旨在解决在线社交平台上的机器人滥用问题,提高在线讨论的完整性。
TwiBot-20 is a large-scale Twitter bot detection benchmark dataset created by Xi'an Jiaotong University. It contains 229,573 users, 33,488,192 tweets, 8,723,736 user attribute items, and 455,958 follow relationships. This dataset was constructed via breadth-first search starting from a large number of seed users across diverse topics, traversing follow relationships to ensure the diversity of users in terms of geographical locations and interest domains, thus better representing the current Twitter ecosystem. TwiBot-20 includes three types of user information modalities, supporting single-user binary classification and community-aware bot detection approaches. Its application domain mainly focuses on Twitter bot detection, aiming to address the issue of bot abuse on online social platforms and improve the integrity of online discussions.

- 1TwiBot-20: A Comprehensive Twitter Bot Detection Benchmark西安交通大学 · 2021年



