Recomendação de conteúdo em um contexto de big data

Abstract

This work uses knowledge extraction in social media based on content recommendation. The process is aided in Big Data aproaches and uses sentiment mining. The growing in social media is a challenge for companies and reasearchers focusing knowledge extraction of massive data in real time. This work evaluated the user preference for technology companies. Sentiment mining is used to measure user preference for a given company. The gathered data are used to build content recommenders and its results are analyzed. The UserBased recommender, with NearestNUser neighborhood and Tanimoto Coefficient similary obtained results closest to ideal, with a score of 0.020, precision of 0.246, recall of 0.259 and F1 measure of 0.252.