A Literature Review in Preprocessing for Sentiment Analysis for Brazilian Portuguese Social Media

Online Social Networks have been increasingly adopted by web users interested in sharing their opinions and thoughts about restaurants, bars, and products they have visited or bought. This scenario enables new analyses to companies and institutions that seek information on how their audience perceives them, and which aspects should be improved. One technique widely used in this type of study is Sentiment Analysis (SA), which allows the automatic mining of opinions on a particular topic. However, this approach faces challenges in social networks, due to the informal nature of the posts and the lack of attention to the grammatical rules found on user-generated content. In this context, this paper presents a literature review about methods and techniques used in the preprocessing of social media data for SA, in the context of Brazilian Portuguese. The results highlight some gaps in the literature and research possibilities, mainly to increase the accuracy of analyses for those platforms.