A Sentiment Aggregation System based on an OWA Operator

Abstract

User-Generated-Content (UGC) in the form of online reviews can be an invaluable source of information for both customers and businesses. Sentiment analysis and opinion mining tools and techniques have been proposed in the literature to extract knowledge from online reviews. Aspect-based opinion mining which has gained growing attention mainly has two tasks including aspect extraction and sentiment polarity detection. Once an aspect-based opinion mining task has been accomplished; a bag of sentiments will be achieved. In many cases, it is necessary to obtain an overall sentiment about a typical aspect. In this study, we have proposed a sentiment aggregation system based on weighted selective aggregated majority OWA (WSAM-OWA). WSAM-OWA considers both the majority and the degree of importance of information source in the process of aggregation. The proposed system exploits the helpfulness rating of reviews in determining the reliability and credibility of each sentiment. A case study was conducted to illustrates the usefulness of the proposed system. The results of this study demonstrated that the proposed sentiment aggregation system could be incorporated in opinion mining systems.