Implementación de un sistema para el análisis de sentimientos en un conjunto de datos

Abstract

In this work is presented the design and implementation of a system to perform sentiment analysis over a dada set. In chapter one, the theoretical framework is established, for which a brief summary of Twitter, the fundamental concepts of sentiment analysis, Hadoop and its components, the Mahout library, and Complement Naïve Bayes Classifier Algorithm, are made. In chapter two, a summary of the design and implementation of the four main components of the system: data collector, data preprocessor, cluster to run Hadoop and analyze data, and results collector is presented. In chapter three, the test of the system are presented, for which five classification models were trained and the one with the best performance was evaluated using additional data; later an analysis of the evaluations’ results was carried out. Finally, in chapter four, the conclusions obtained and the recommendations that could contribute to future works are presented.