Analyzing tie-strength across different media

Human interactions are becoming more and more important in computer systems. The interactions can be classified according to predefined rules, based on the assumption that relations between humans differ greatly. The tie-strength notion is used throughout the social sciences literature in order to denote this classification. The present paper researches how the tie-strength between two persons can be computed automatically by applying structural data from different sources, e.g. email, shared workspaces. This data can provide a virtual copy of the ego-centric network of a user and therefore be utilized for social intelligent computing.