Employing clustering algorithms to create user groups for personalized context aware services provision

Abstract

The successful provision of context aware services entails the attainment of equilibrium between the extent of personalization desired and the user's need for privacy. Two are the major elements that play a significant role: the user's location and the user's preferences. In this paper we focus on the latter, and propose to employ a social groups' creation methodology, so as to hierarchically organize the user preferences concerning any domain in different levels of detail. We describe some notions and metrics which play a key role in social networking frameworks, and we perform an evaluation study of three widely used clustering methods (k-means, hierarchical and spectral clustering) in the scope of social groups assessment and in regard to the cardinality of the profile used to assess users' preferences. The results of the work can be used in many applications, including personalized media delivery, offering a framework on which next generation multimedia access can be provided.