Sentiment Polarity Classification at EVALITA: Lessons Learned and Open Challenges

Sentiment analysis in social media is a popular task attracting the interest of research community, also recent evaluation campaigns natural language processing tasks several languages. We report on our experience organization SENTIment POLarity Classification Task (SENTIPOLC), shared sentiment classification Italian tweets, proposed for first time 2014 within Evalita campaign. present datasets-which include an enriched annotation scheme dealing with impact figurative polarity-the methodology, and discuss approaches results participating systems. offer reflection open challenges state-of-the-art systems microblogging Italian, as they emerge from qualitative misclassified tweets. Finally, we provide resources have created, share lessons learned by running this two consecutive editions.