Identificando emoções em manchetes de notícias escritas em português do Brasil utilizando Naïve Bayes

Abstract

The large increase in the number of users on the internet has allowed that data contains opinion or sentiment of user to grow at an amazing speed. In this way, most newspapers offer their daily publications on their sites. Considering the popularization on the internet, this big amount of available data on the internet grows exponentially and analyzing such data in a non-automated way can be a non-trivial problem. Thus, this paper shows a method to identify automatically emotions in text. The goal is to build a model able to identify one of the six basic emotions (joy, sadness, surprise, fear, anger and disgust) in news headlines written in Brazilian Portuguese. For the build of the model, we use a machine learning approach based on the Naïve Bayes classifier. The corpus (dataset) used in this work is composed by news extracted from online newspapers. This news were processed by our Naïve Bayes classifier getting an accuracy rate of 59,5%.