Improving Topic Modeling for Studying Political Text

• In Event: Text as Data

Brief Overview

We study Structural Topic Models and Joint Sentiment Topic models which offer insights into politicians' communications. Comparing both to LDA, we find that JST produces more coherent topics and better identifies partisanship across domains.

Authors

• Daniel Ebanks, California Institute of Technology

• Michael Alvarez, California Institute of Technology

• Nicholas J Adams-Cohen, Stanford University