Stance and Sentiment in Czech

Abstract

Sentiment analysis is a wide area with greatpotential and many research directions. One direction isstance detection, which is somewhat similar to sentimentanalysis. We supplement stance detection dataset withsentiment annotation and explore the similarities of thesetasks. We show that stance detection and sentimentanalysis can be mutually beneficial by using gold labelfor one task as features for the other task. We analysedthe presence of target entities for stance detection in thedataset. We outperform the state-of-the-art results forstance detection in Czech and set new state-of-the-artresults for the newly created sentiment analysis part ofthe extended dataset.