Diversifying Product Review Rankings: Getting the Full Picture
2011 (English)In: 2011 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology, Washington DC: IEEE Computer Society Digital Library, 2011, 138-145 p.Conference paper (Refereed)
E-commerce Web sites owe much of their popularityto consumer reviews provided together with product descriptions.On-line customers spend hours and hours going through heaps oftextual reviews to build confidence in products they are planningto buy. At the same time, popular products have thousands ofuser-generated reviews. Current approaches to present them tothe user or recommend an individual review for a product arebased on the helpfulness or usefulness of each review. In thispaper we look at the top-k reviews in a ranking to give a goodsummary to the user with each review complementing the others.To this end we use Latent Dirichlet Allocation to detect latenttopics within reviews and make use of the assigned star ratingfor the product as an indicator of the polarity expressed towardsthe product and the latent topics within the review. We present aframework to cover different ranking strategies based on theuser’s need: Summarizing all reviews; focus on a particularlatent topic; or focus on positive, negative or neutral aspects.We evaluated the system using manually annotated review datafrom a commercial review Web site.
Place, publisher, year, edition, pages
Washington DC: IEEE Computer Society Digital Library, 2011. 138-145 p.
ranking; diversification; review recommendation; summarizationtion; topic models
Computer and Information Science
IdentifiersURN: urn:nbn:se:kth:diva-84406DOI: 10.1109/WI-IAT.2011.33ScopusID: 2-s2.0-80155182260ISBN: 978-1-4577-1373-6OAI: oai:DiVA.org:kth-84406DiVA: diva2:499653
2011 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology
Winner of best paper award at 2011 IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology.
© 2011 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
QC 201202152012-02-152012-02-132012-02-15Bibliographically approved