Analyzing Framing through the Casts of Characters in the News

Dallas Card1, Justin Gross2, Amber Boydstun3, Noah A. Smith4
1Carnegie Mellon University, 2University of Massachusetts, Amherst, 3University of California, Davis, 4University of Washington


Abstract

We present an unsupervised model for the discovery and clustering of latent "personas" (characterizations of entities). Our model simultaneously clusters documents featuring similar collections of personas. We evaluate this model on a collection of news articles about immigration, showing that personas help predict the coarse-grained framing annotations in the Media Frames Corpus. We also introduce automated model selection as a fair and robust form of evaluation.