Entity-centric Sentiment Analysis on Twitter data for the Potuguese Language

摘要

Twitter is a popular microblogging platform which is commonly used to express opinions about entities of the world. The solutions provided to perform Sentiment Analysis in such a media, however, relies on classifying an entire sentence regarding the opinion it express, rather than the content and reference of the opinion expressed in the text. We propose and evaluate a Entity-centric Sentiment Analysis method over Twitter data for the Portuguese language.