Soft rumor control in social networks: Modeling and analysis

Nowadays, social networks become ubiquitous platforms for sharing and diffusing information around the world. However, spreading rumors as unverified and opaque information in social networks causes harmful damages to societies. An approach for combating rumors in social networks is to use soft control mechanisms i.e. enhancing the people?s knowledge and awareness against the rumor to persuade them avoiding rumor dissemination. In this paper, we propose a soft rumor control model in which people refer to their trusted friends or ask the reputable authorities about the rumor to avoid rumor spreading. The model includes a method for selecting consultants who are both expert in rumor context and responsive to queries about rumors by the user. The battlespace between rumor and anti-rumor spreaders is then modeled as an evolutionary game to analyze the controls? effectiveness. To evaluate the proposed model, we use Pheme dataset of tweets and conduct simulation analysis. It is shown that trusted consultants suggested by the model with high precision are the same users who send anti-rumor messages in real world. Furthermore, we analyze and compare soft rumor control methods on societies with different assumed cyber literacy and habits. Moreover, it is interestingly shown that using soft rumor control mechanisms in some situations outperforms traditional hard controls (e.g. censorship). Note that as we have used tangible factors in formulating the proposed model, it can help social network developers to build feasible soft rumor control facilities in their own products.