Opinion mining in social media: Modeling, simulating, and forecasting political opinions in the web

Affordable and ubiquitous online communications (social media) provide the means for flows of ideas and opinions and play an increasing role for the transformation and cohesion of society – yet little is understood about how online opinions emerge, diffuse, and gain momentum. To address this problem, an opinion formation framework based on content analysis of social media and sociophysical system modeling is proposed. Based on prior research and own projects, three building blocks of online opinion tracking and simulation are described: (1) automated topic, emotion and opinion detection in real-time, (2) information flow modeling and agent-based simulation, and (3) modeling of opinion networks, including special social and psychological circumstances, such as the influence of emotions, media and leaders, changing social networks etc. Finally, three application scenarios are presented to illustrate the framework and motivate further research.