Social Media vs. News Media: Analyzing Real-World Events from Different Perspectives

摘要

For a long time, the news media has played a crucial role not only as an information provider, but also as an influential source of opinion and commentary. Nowadays, platforms such as Twitter provide an alternative to the traditional one-way interaction, enabling users to voice their opinions. Hence, one can obtain a more comprehensive picture of the range of perspectives on real-world events by considering both news and social media sources. In this paper, we compare mainstream news and Twitter data on 18 well-known real-world events from six different categories. We propose the event-based authoring model (EvA), a novel probabilistic model to capture the content characteristics of an event with respect to aspect, category and background word distributions. These results allow us to analyze the real-world events in different perspectives.