A percepção dos usuários sobre a utilização da análise de sentimentos como apoio à seleção de notícias

Everyday, readers are bombarded with various types of news (positive or negative) without having the chance to choose what they want to read. In this paper we investigate the use of sentiment analysis to support the user in the content selection. We present Magnet News, a Web tool where readers can choose the polarity of the news they want to read. In this scenario we use the Underlying Discourse Unveiling Method (UDUM) to evaluate the user experience in relation to the proposal. Our results show that the sentiment analysis can be an important mechanism to support users in the news choice. Author