Unsupervised Opinion Polarity Detection based on New Lexical Resources

Abstract

There are polarity detection techniques based on the lexicon of opinion words and those based on machine learning techniques. In this paper, we focus on unsupervised polarity detection using lexical resources. We present the SentiWordNet 4.0 and the SpanishSentiWordNet in order to solve the detected drawbacks of previous resources. The integration of the proposed resources is solved by combining them in the PolarityDetection library, which is integrated to PosNeg Opinion 2.0 and facilitates obtaining high accuracy and recall values.