北京大学学报（自然科学版） ›› 2025, Vol. 61 ›› Issue (4): 639-649.DOI: 10.13209/j.0479-8023.2024.120

上一篇    下一篇

基于社交时序知识图谱的推特机器人检测方法

蒋致书, 陈炜, 张伟杰, 张诗琪, 陈季若, 万怀宇†   

  1. 北京交通大学计算机与信息技术学院, 北京 100044
  • 收稿日期:2024-05-30 修回日期:2024-07-25 出版日期:2025-07-20 发布日期:2025-07-20
  • 通讯作者: 万怀宇, E-mail: hywan(at)bjtu.edu.cn
  • 基金资助:
    国家重点研发计划(2021QY1502)资助

Twitter Bot Detection Method Based on Social Temporal Knowledge Graph

JIANG Zhishu, CHEN Wei, ZHANG Weijie, ZHANG Shiqi, CHEN Jiruo, WAN Huaiyu†   

  1. School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044
  • Received:2024-05-30 Revised:2024-07-25 Online:2025-07-20 Published:2025-07-20
  • Contact: WAN Huaiyu, E-mail: hywan(at)bjtu.edu.cn

PDF

摘要：

为了解决现有推特机器人检测方法对用户动态社交历史结构性和时序性的忽视以及特征融合引起的噪声累积问题, 构建社交时序知识图谱STKG, 并提出一种推特机器人检测方法STKGBot。在STKG中, STKGBot使用关系增强图注意网络RE-GAT来学习静态社交关系特征, 使用时序增强图卷积网络TE-GCN来学习动态社交历史特征, 使用双线性模型进行特征融合。此外, STKGBot使用对比学习来缓解特征融合引起的噪声累积。在两个公开数据集上的实验结果表明, STKGBot的检测结果均优于当前最先进的模型。

关键词: 推特机器人检测, 时序知识图谱, 图神经网络, 对比学习

Abstract:

Existing Twitter bot detection methods often overlook the structural and temporal information of users’ dynamic social history, as well as the noise accumulation resulting from feature fusion. In order to address these limitations, this paper constructs STKG (social temporal knowledge graph) and proposes a Twitter bot detection method STKGBot (STKG for Twitter bot detection). In the STKG, STKGBot uses RE-GAT (heterogeneity-enhanced graph attention network) to learn the static social relationship feature, TE-GCN (temporal-enhanced graph convolutional network) to learn the dynamic social history feature, and a bilinear model for the feature fusion. In addition, STKGBot employs contrastive learning to alleviate the noise aggravation in the process of feature fusion. Experimental results on two public datasets demonstrate that STKGBot outperforms state-of-the-art models.

Key words: Twitter bot detection, temporal knowledge graph, GNN, contrastive learning

版权所有 © 《北京大学学报（自然科学版）》编辑部
技术支持：北京玛格泰克科技发展有限公司