Mineração de textos: análise de sentimentos utilizando Tweets referentes à Copa do Mundo 2014

Abstract

The increase of the social networks in the last years allowed users to get connected and share information in real time, spreading it for thousands of others users in a short time. Also, the way that users interact has changed. The users of social networks usually post their opinions about big events, product releases, catastrophes, epidemics, among other happenings. To follow what they are talking on the social networks may be a differential for organizations who wants to elaborate better marketing strategies, to obtain feedback about some product or a certain event, among other possibilities. Although, this big amount of data still keeps growing, and analysis it in a non-automated way may be a non-trivial problem. In this context, this article shows how the Text Mining process was used to collect, to structure the text extracted from Twitter(tweets) and to create a text classification model that allowed to predict the Twitter social network user‟s opinion about the FIFA World Cup Brazil 2014. The user‟s posts called as tweets are categorized in this work as a sentiment: positive, negative, ambiguous or neutral.