Towards a Framework for Contextualising Potential Fake News on Twitter (Master Thesis, Finished)
In recent years, the increasing spread of false or misleading information, also called "Fake News", has had real world effects, for example on the outcome of the 2016 U.S. presidential election. This is an immensely relevant subject due to the risk of being abused for propaganda purposes.
For this Master's Thesis, we propose a framework that helps a user in asserting the validity of Tweets found in a continuous stream of Twitter. Our framework foresees a three-step approach: In a first step, we will filter out Tweets, that we cannot analyze, e.g. Tweets that are too short. Subsequently, we plan to analyze social network graph information extracted from Twitter as well as external evidence from search engines, e.g. via the Custom Search JSON API by Google. This analysis will give us partial verdicts, which we then combine and contextualize in a final step, before they are returned to the user, to help in asserting the truthfulness and the reliability of the Tweet and its content.
Start / End Dates
2019/06/05 - 2019/12/04