Abstract-While the analysis of online social networks is a prominent research topic, offline real-world networks are still not covered extensively. However, their analysis can provide important insights into human behavior. In this paper, we analyze influence factors for link prediction in human contact networks. Specifically, we consider the prediction of new links, and extend it to the analysis of recurring links. Furthermore, we consider the impact of stronger ties for the prediction. The results and insights of the analysis are a first step onto predictability applications for human contact networks.
I. INTRODUCTIONWith the growing amount of social data, ubiquitous systems, and mobile social media applications transcending everyday life, the analysis of social networks is receiving increased attention. This especially relates to the dynamics and the creation of links between the networks' subjects [20], important aspects of offline social networking still remains largely unexplored. The analysis of such networks can potentially provide more direct answers to fundamental questions, e. g., how do personal links get established, is it possible to correlate this with roles, how does the intensity of personal communication evolve?In this paper, we aim at providing first insights for answering such questions. We focus on real-world offline networks of human contacts, that is, face-to-face conversations between persons. In contrast to virtual networks, these contacts were acquired using a ubiquitous RFID-based system that allows us to collect face-to-face contacts. Thus, we can observe and analyze social interaction at a very detailed level, including the specific event sequences and durations.We [16] in the context of networks of human contacts. We aim to predict new contacts based on network properties, as an adaptation of methods for online social networks. In addition, we extend the analysis in two important directions: First, we consider the length of the contacts in more detail, and analyze the impact of longer conversations. Second, we consider the prediction of future recurring contacts, i. e., renewed contacts between specific actors. For these, we analyze influence factors and patterns for establishing such contacts, and also consider their specific durations in a fine-grained dynamic analysis. Essentially, this leads to the analysis of the impact of stronger ties for new and recurring contacts.