False Information Detection on Social Media via a Hybrid Deep Model

There is not only low-cost, easy-access, real-time and valuable information on social media, but also a large amount of false information. False information causes great harm to individuals, the society and the country. So how to detect false information? In the paper, we analyze false information further. We rationally select three information evaluation metrics to distinguish false information. We pioneer the division of information into 5 types and introduce them in detail from the definition, the focus, features, etc. Moreover, in this work, we propose a hybrid deep model to represent text semantics of information with context and capture sentiment semantics features for false information detection. Finally, we apply the model to a benchmark dataset and a Weibo dataset, which shows the model is well-performed.