A Framework towards Computational Narrative Analysis on Blogs

Social media is widely used to express views and share opinions with others. With the availability of inexpensive and ubiquitous mass communication tools like social media, creating narratives, falseinformation and propaganda is both convenient and e↵ective. Social media users leverage this platform to further their views by framing narratives and participating in online discourse. Almost all events, issues, crises are discussed on social media. Blogs, unlike other social media platforms, are not regulated by any authority and have no restriction on character limit, providing bloggers with not only space for richer content but also serve as a platform for agenda-setting and content framing abetting development of narratives. This innate feature of blogs makes them a valuable platform for sociologists/political scientists to gain situational awareness by tracking di↵erent opinions, political views, and narratives as they are shaped. However, the deluge of posts in blogosphere makes it impractical to manually identify narratives. In this paper, we propose a novel framework to computationally identify narratives by extracting actors/actions using NLP techniques including POS tagging, chunking, and grammar rules to identify narratives. We also employ a scoring mechanism to rank them in the order of their dominance. Later, we evaluate the e cacy of the proposed model by validating against human annotated narratives. Our framework achieved an accuracy of 66.8%. Our proposed framework can help social scientists identify narratives computationally reducing human e↵ort reasonably. Moreover, the results from this research could also be used to build e↵ective counter narratives to stem propaganda campaigns.