Research on Weibo Misinformation Detection Based on Heterogeneous Information Network Representation Learning
[Research purpose]With the rapid development of social networks and the occurrence of public health emergencies,a large a-mount of deceptive misinformation is often intertwined within social media platforms.Identifying such misinformation has become an im-portant component of users'information literacy.[Research method]This study focuses on heterogeneous information networks on Wei-bo.By considering both the semantic features of Weibo texts and the non-semantic features of user behaviors,a multi-head attention mechanism is introduced to integrate generated ensemble representations for misinformation detection.Feature mining and quantitative anal-ysis are conducted from three dimensions:information content,participating users,and user-information interactions.[Research conclu-sion]Research indicates that the misinformation detection method based on heterogeneous information network representation learning demonstrates practicality.It contributes to deconstructing the characteristics of misinformation,thereby providing beneficial insights for governing and strategizing against misinformation during public health emergencies.
heterogeneous information networkrepresentation learningpublic health emergenciesmisinformationuser behaviorso-cial mediaWeibo text
万方数据
维普
