An analysis of slant in tweets: Case study

Abstract

Determination of quality and reliability of information found in social media have been subjects of study by sever researchers. One set of solution may not work in all cases. This paper presents a method to estimate the slant of tweets related to a topic. The general approach followed is to construct labeled data from tweets and use supervised learning to build predictive models. Results obtained from two datasets are compared against OTC model and a CNN based model.