Contextual Social Networking

The thesis centers around the multi-faceted research question of how contexts may be detected and derived that can be used for new context aware Social Networking services and for improving the usefulness of existing Social Networking services, giving rise to the notion of Contextual Social Networking. In a first foundational part, we characterize the closely related fields of Contextual-, Mobile-, and Decentralized Social Networking using different methods and focusing on different detailed aspects. A second part focuses on the question of how short-term and long-term social contexts as especially interesting forms of context for Social Networking may be derived. We focus on NLP based methods for the characterization of social relations as a typical form of long-term social contexts and on Mobile Social Signal Processing methods for deriving short-term social contexts on the basis of geometry of interaction and audio. We furthermore investigate, how personal social agents may combine such social context elements on various levels of abstraction. The third part discusses new and improved context aware Social Networking service concepts. We investigate special forms of awareness services, new forms of social information retrieval, social recommender systems, context aware privacy concepts and services and platforms supporting Open Innovation and creative processes. This version of the thesis does not contain the included publications because of copyrights of the journals etc. Contact in terms of the version with all included publications: Georg Groh, grohg@in.tum.de.