Investigating the Usefulness of Product Reviews Through Bipolar Argumentation Frameworks

The importance of useful product reviews cannot be overstated, as they not only provide crucial information to potential buyers but also offer valuable feedback to the businesses or individuals under review. Providing useful reviews to consumers has gained significant attention as an effective market analysis tool for companies. Numerous studies have delved into the facets that contribute to predicting the usefulness of reviews, encompassing a spectrum from philosophical insights to advancements in artificial intelligence. In this work, we study how to use the argument structure of reviews to identify the most useful reviews in a set of annotated product reviews. In particular, we use quantitative bipolar argumentation frameworks (QBAFs) to construct a model of review arguments and topics, and we apply reasoning to such a model to identify useful reviews. Our results show that indeed argument reasoning, and QBAFs in particular, provide an insightful and well-performing means to analyze the usefulness of reviews.