Contextual and Personalized Mobile Recommendation Systems

Abstract

Mobile devices are becoming ubiquitous. People are getting used to using their phones as a personal concierge to discover what is around and decide what to do. Mobile recommendation therefore becomes important to understand user intent and simplify task completion on the go. Since user intents essentially vary with users and sensor contexts (time and geo-location, for example), mobile recommendation needs to be both contextual and personalized. While rich user mobile data is available, such as mobile query, click-through, and check-in record, there exist two challenges in utilizing them to design a contextual and personalized mobile recommendation system: exploring characteristics from large-scale and heterogeneous mobile data and employing the uncovered characteristics for recommendation. In this chapter, the authors talk about two mobile recommendation techniques that well address the two challenges. (1) One exploits mobile query data for local business recommendation, and (2) one exploits mobile check-in record to assist activity planning.