Link Mining and Localisation in the Context of Face-to-Face Contact Networks

For a long time an automatic detection of contacts between humans was not possible. In this work a new generation of resource-aware RFID tags (proximity tags) is used which has the ability to detect reliable face-to-face contacts. This innovation opens up new research possibilities in the fields of human contact behaviour analysis, link prediction and indoor localisation.

In this context the identification of human contact structures and their underlying pro¬cesses is a prominent research topic. However, the analysis of offline social networks has been largely neglected. In this work face-to-face information is utilised to study the link prediction problem as well as dynamic and static contact patterns in face-to-face contact networks. Furthermore the influence of user interests and social contacts on the predictability of talk attendance is analysed.

The localisation of humans in indoor environments is still a challenging problem. In literature, accurate positioning approaches exist. Unfortunately, these approaches requi¬re expensive hardware and an extensive deployment of suitable infrastructure. Therefore novel approaches are presented that use proximity tags for positioning. All methods are evaluated using real-world data and it is shown that all approaches significantly outperform state-of-the art indoor localisation approaches.