The CLSA model: A novel framework for concept-level sentiment analysis

Abstract

Hitherto, sentiment analysis has been mainly based on algorithms relyingon the textual representation of online reviews and microblogging posts.Such algorithms are very good at retrieving texts, splitting them into parts, checkingthe spelling, and counting their words. But when it comes to interpretingsentences and extracting opinionated information, their capabilities are knownto be very limited. Current approaches to sentiment analysis are mainly basedon supervised techniques relying on manually labeled samples, such as movieor product reviews, where the overall positive or negative attitude was explicitlyindicated. However, opinions do not occur only at document-level, nor theyare limited to a single valence or target. Contrary or complementary attitudes towardthe same topic or multiple topics can be present across the span of a review.In order to overcome this and many other issues related to sentiment analysis,we propose a novel framework, termed concept-level sentiment analysis (CLSA)model, which takes into account all the natural-language-processing tasks necessaryfor extracting opinionated information from text, namely: microtext analysis,semantic parsing, subjectivity detection, anaphora resolution, sarcasm detection,topic spotting, aspect extraction, and polarity detection.