Análisis de sentimientos para textos cortos en español, una revisión del estado del arte

Abstract

In general, social networks are currently used by users for communication, comments, and opinions. So, the automatic analysis of them on topics of interest requires an exhaustive review by machines. With sentiment analysis, it is possible to determine what users want to express. Currently, many tools allow this sentiment analysis to be carried out for short texts. However, most of them are focused on the English language. This article aims to study the tools and Corpus used to analyze sentiments of texts in Spanish through state of art. Among the most important results, it was found that the most used Corpus is the TASS. In addition, a comparison was made between the methods used, including SentiWordNet, Bayes, iSol, the most efficient being SVM.