Exploiting Feedback from Users of innsbruck.mobile for Personalization

Abstract

innsbruck.mobile is a mobile tourist guide for visitors of the city ofInnsbruck. The system allows the retrieval of detailed information aboutsights, accommodation, events, dining out and weather forecasts as wellas pushes recommendations from these categories to registered visitorsvia a short-messaging service. Personalising the information access andfiltering of content according to the assumed relevance to its users wasa key requirement for the system especially in the context of mobileservices with limited display capabilities. However, most real usersconduct few and very short interactions with the system and thereforeprovide only sparse relevance feedback. As a consequence collaborativerecommendation methods need to utilise multiple sources of user feedbackto increase the accuracy on sparse rating data. We give a description ofour system and conduct an evaluation based on this sparse usage datathat includes more than 800 distinct user sessions.