Assessing the Helpfulness of Review Content for Explaining Recommendations
This addresses the need for transparency in recommender systems for users, but it is incremental as it adapts existing argumentation theories to a new domain.
The paper tackled the problem of generating explanations for recommender systems by identifying helpful content in reviews, finding that credibility and convincingness mediate the relationship between helpfulness and objectivity/relevance.
Despite the maturity already achieved by recommender systems algorithms, little is known about how to obtain and provide users with a proper rationale for a recommendation. Transparency and effectiveness of recommender systems may be increased when explanations are provided. In particular, identifying of helpful argumentative content from reviews can be leveraged to generate textual explanations. In this paper, we investigate the reasons why a review might be considered helpful, and show that the perception of credibility and convincingness mediates the relationship between helpfulness and the perception of objectivity and relevant aspects addressed. Our findings led us to suggest an argumentbased approach to automatically extracting helpful content from hotel reviews, a domain that differs from those that best fit classical argumentation theories.