A Recommender System for Online Consumer Reviews

Online consumer reviews have helped consumers to increase their knowledge about different products/services. While most previous studies try to provide general models that predict performance of online reviews, this study notes that different people look for different types of reviews. Hence, there is a need for developing a system that that is able to sort reviews differently for each user based on the ratings they previously assigned to other reviews. Using a design science approach, we address the above need by developing a recommender system that is able to predict the perceptions of each user regarding helpfulness of a specific review. In addition to addressing the sorting problem, this study also develops models that extract objective information from the text of online reviews including utilitarian cues, hedonic cues, product quality, service quality, price, and product comparison. Each of these characteristics may also be used for sorting and filtering online reviews.