Identifying Suspected Cybermob on Tieba

摘要

This paper describes an approach to identify suspected cybermob on social media. Many researches involve making predictions of group emotion on Internet (such as quantifying sentiment polarity), but this paper instead focuses on the origin of information diffusion, namely back to its makers and contributors. According our previous findings that have shown, at the level of Tieba's contents, the negative information or emotions spread faster than positive ones, we centre on the maker of negative message in this paper, so-called cybermobs who post aggressive, provocative or insulting remarks on social websites. We explore the different characteristics between suspected cybermobs and general netizens and then extract relative unique features of suspected cybermobs. We construct real system to identify suspected cybermob automatically using machine learning method with above features, including other common features like user/content-based ones. Empirical results show that our approach can detect suspected cybermob correctly and efficiently as we evaluate it with benchmark models, and apply it to actual cases.