Um dataset para análise de sentimentos na língua portuguesa

Abstract

Social networks have become important sources of information in which shared content can tell a lot about their users' opinions and feelings on a variety of subjects, among them, Twitter stands out as one of the most popular networks. Analyze the feelings exposed by the social network users can help understand what people are talking about a particular company, brand, event or even other people, functioning as a way to get feedback. This article presents the creation of a dataset for sentiment analysis, containing data collected from public messages in Portuguese on Twitter, its usage, and evaluation.