Tie Strength in Co-authorship Social Networks: Analyses, Metrics and a New Computational Model

Abstract

The study of social ties has lead to build rigorous models that reveal the evolution of social networks and their dynamism. A property related to social ties is the strength of ties, which has been largely explored in different contexts, such as information diffusion, analyses of patterns in communication logs and evaluation of scientific researchers productivity. Specially, analyzing tie strength allows investigating how distinct relationships play different roles and identifying impact at micro-macro levels in the network. We present and propose different ways to measure the strength of co-authorship ties in non-temporal and temporal real academic social networks. Specially, tie strength can be measured by topological and semantic properties, as well as their combination. Finally, this thesis reveals different concepts that define tie strength and properties that influence it, along with metrics, algorithms and a classification for distinct relationships.