A Hybrid Method for Sentiment Analysis of Election Related Tweets

Political sentiment analysis using social media content has recently attracted significant interest. This paper focuses on analyzing tweets in Greek regarding the recent European Elections. A hybrid method that combines Greek lexicons and classification methods is presented. A probabilistic classification model, that predicts the sentiment of tweets, is combined with hashtag-based filtering. Based on the predictions, an analysis is implemented, that shows how the public sentiment was affected by specific events during the pre-election period.