Fine-Grained Sentiment Analysis for Movie Reviews in Bulgarian

We present a system for fine-grained sentiment analysis in Bulgarian movie reviews. As this is pioneering work for this combination of language and sentiment granularity, we create suitable, freely available resources: a dataset of movie reviews with fine-grained scores, and a sentiment polarity lexicon. We further compare experimentally the performance of classification, regression and ordinal regression in a 3-way, 5-way and 11-way classification setups, using as features not only the text from the reviews, but also contextual information in the form of metadata, e.g., movie length, director, actors, genre, country, and various scores: IMDB, Cinexio, and user-average. The results show that adding contextual information yields strong performance gains.