Mineração de textos do Twitter utilizando técnicas de classificação

Abstract

In recent years, due to the growth of social networks, people interact and share relevant information on the Internet. This growth has some advantages. One of the advantages is the sharing of information between users in a short time. Some social networks users often post reviews of different events that are the breaking news in that moment. Consequently, the daily volume of data that is generated through social networks grows exponentially and all the information generated through the may be relevant if it is treated and used correctly. Therefore, there is the stimulus to generate knowledge from this information in an organized and automated. The main purpose of this work is to construct a model to classify into categories, the data that was captured from the social network Twitter. Posts of users, known as tweets, are categorized in this work as: economy, sport, religion, politics and others. The data collected from Twitter went through a natural language processing before building the model, in order to remove stopwords, accents, punctuation and suffixes. The obtained model showed satisfactory accuracy and was generated using the Naive Bayes classification algorithm.