Abstract

AMORES, Mario; ARCO, Leticia  and  ARTILES, Michel. PosNeg opinion: A tool for managing comments from the web. Rev cuba cienc informat [online]. 2015, vol.9, n.1, pp.20-12. ISSN 2227-1899.

One of the primary tasks of opinion mining is the classification of the polarity of the opinion, that consists in determining whether the opinion is positive or negative with respect to the entity to which it is referring. This work seeks design a general scheme consists in five stages that enable the polarity detection of opinions in an unsupervised manner. The five considered stages are: token identification, lexical disambiguation, get all meanings of each word, classify each token in positive or negative, and evaluate the opinion. PosNeg Opinion detects the polarity of the opinions in Spanish language. The proposed system showed good performance in classifying the polarity of 200 opinions from discussion forums Yahoo.es, yielding an accuracy of 0.965 and a precision of 0.970.

Keywords : opinion mining; polarity detection; semantic Web; sentiment analysis.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )