Evaluating the effectiveness of stereotype user models for recommendations on mobile devices

Mobile recommender systems have been proven as a promising approach in mobile scenarios to support the decision making process of users by suggesting beneficial items in a certain mobile context. The main goal of this paper is to examine whether a stereotype user model leads to better recommendations as part of such a system. For this purpose, we developed and tested a prototype for a shopping scenario. Research on fashion stereotypes led to a user model containing ten different stereotypes. The stereotype classification is performed by computing the proximity of each stereotype to the user's properties. Results of a user study show that a user model based on stereotypes generates better results than a recommender system without a stereotype-based user model. Moreover, stereotype-based user models allow personalized recommendations right away thus contributing to alleviating the cold start problem.