基于超图理论的多模态社交媒体谣言检测模型研究
首发时间:2023-03-03
摘要:多模态假信息相比传统文本信息传播得更快,会造成了更大的社会危害。因此,多模态谣言检测面临着很多挑战。本研究旨在以社交媒体中帖子内容为研究对象,利用深度神经网络提取信息的文本和视觉特征,再基于超图网络,对不同特征进行融合并根据融合结果预测信息的真假。从而实现社交媒体谣言的早期发现。实验结果表明,本研究提出的方法能够在微博数据集上相比基线方法在准确率等性能指标上有所提升。
For information in English, please click here
Research on rumor detection model with multimodal social media based on hypergraph theory
Abstract:Compared with traditional text information,multimodal false information spreads faster, which will cause greater social harm. Therefore, multimodal rumor detection faces many challenges. The purpose of this study is to take the post content in social media as the research object, extract the text and visual features of information using deep neural network, and then fuse different modal\'s feature based on hypergraph network, and predict the credibility of information according to the fusion results. The experimental results show that the method proposed in this study can improve the accuracy and other performance indicators compared with the baseline method on the microblog data set.
Keywords: Rumor detection Hypergraph Multimodal Social media
基金:
引用
No.****
同行评议
勘误表
基于超图理论的多模态社交媒体谣言检测模型研究
评论
全部评论