Characterizing and predicting fake news spreaders in social networks
Characterizing and predicting fake news spreaders in social networks
复制标题
DOI：
10.1007/s41060-021-00291-z
复制
发表时间：
2021-11-22
影响因子：
2.4
通讯作者：
Spezzano, Francesca
中科院分区：
文献类型：
--
作者：
Shrestha, Anu;Spezzano, Francesca
由于假新闻在社交媒体上的迅速传播以及公众舆论变化的社会威胁，假新闻已经引起了广泛的关注。随着社交媒体对日常新闻的日益普及，用户在传播假新闻方面所扮演的角色已成为不可避免的。社交媒体中的人们积极参与新闻的创作和传播，有意无意地助长了假新闻的扩散。因此，有必要识别那些倾向于分享假新闻的用户，以减少假新闻在社交媒体上的猖獗传播。在本文中，我们对从 Twitter 收集的两个不同数据集进行了全面分析，并研究了存在错误信息的社交媒体中的用户特征模式。具体来说，我们研究了用户特征与其成为假新闻传播者的可能性之间的相关性，并证明了所提出的特征在识别假新闻传播者方面的潜力。我们提出的方法在所考虑的数据集上实现了 0.80 到 0.99 之间的平均精度，始终优于基线模型。此外，我们还表明，用户的个性特征、情绪和写作风格是假新闻传播者的有力预测因素。
Due to its rapid spread over social media and the societal threat of changing public opinion, fake news has gained massive attention. Users' role in disseminating fake news has become inevitable with the increase in popularity of social media for daily news diet. People in social media actively participate in the creation and propagation of news, favoring the proliferation of fake news intentionally or unintentionally. Thus, it is necessary to identify the users who tend to share fake news to mitigate the rampant dissemination of fake news over social media. In this article, we perform a comprehensive analysis on two different datasets collected from Twitter and investigate the patterns of user characteristics in social media in the presence of misinformation. Specifically, we study the correlation between the user characteristics and their likelihood of being fake news spreaders and demonstrate the potential of the proposed features in identifying fake news spreaders. Our proposed approach achieves an average precision ranging between 0.80 and 0.99 on the considered datasets, consistently outperforming baseline models. Furthermore, we also show that the user personality traits, emotions, and writing style are strong predictors of fake news spreaders.