Generación de recursos para Análisis de Opiniones en español

Abstract

Sentiment Analysis (SA) refers to the treatment of the subjective information in texts, product reviews, comments on blogs or personal opinions. One of the basic tasks in SA is classifying the polarity of a given text in a document, i.e., whether the opinion expressed is positive, negative, or neutral. Many studies have investigated the polarity classification in documents written in English. However, nowadays more and more people express their comments, opinions or points of view in their own language. For this reason, it is necessary to develop systems than can extract and analyze all this information in different languages. In this work we focus on polarity detection for Spanish reviews. We are mainly concerned with linguistic resources for Spanish sentiment analysis because, in addition to the lack of resources for this language in this area, it is currently the third most used language in the web 2.0.