Ranking reader emotions using pairwise loss minimization and emotional distribution regression

Abstract

This paper presents two approaches to ranking reader emotions of documents. Past studies assign a document to a single emotion category, so their methods cannot be applied directly to the emotion ranking problem. Furthermore, whereas previous research analyzes emotions from the writer's perspective, this work examines readers' emotional states. The first approach proposed in this paper minimizes pairwise ranking errors. In the second approach, regression is used to model emotional distributions. Experiment results show that the regression method is more effective at identifying the most popular emotion, but the pairwise loss minimization method produces ranked lists of emotions that have better correlations with the correct lists. © 2008 Association for Computational Linguistics.