A study of Media Polarization with Stylometry Methods
Keywords:Natural Language Processing, Media Polarization, Authorship Attribution, Stylometry
This research investigated the U.S. media polarization with stylometry approaches, creating classification models to identify the political leanings of news articles based on their writing style. We tested the models of authorship attribution, while controlling for topic, stance, and style, and applied them to media companies and their identity within a political spectrum. We tested style features that could include semantic and/or sentiment-related information, such as stance taking, with features that seemingly do not capture it. We were able to successfully classify articles as left-leaning or right-learning regardless of stance. Finally, we provide an analysis of some of the patterns that we found.