A Multi Ego-Network Model for Synthetic Social Network Graphs

Human so ial relationships are a key omponent of emerging complex te hno-soc ial systems su h as soc ially- entri platforms based on the intera tions between humans and ICT te hnologies. Therefore, the models of human so ial relationships are fundamental to hara terise these systems and study the performan e of so ially- entri platforms depending on the so ial ontext where they operate. The goal of this pa- per is presenting a generative model for building syntheti human so ial network graphs where the properties of so ial relationships are a urately reprodu ed. The model goes well beyond a binary approa h, whereby edges between nodes, if existing, are all of the same type. It sets the properties of ea h so ial link, by in orporating fundamental results from the an- thropology literature. The syntheti networks it generates a urately re- produ e both the ma ros opi stru ture (e.g., its diameter and luster of oeÆ ient), and the mi ros opi stru ture (e.g., the properties of the tie strength of individual so ial links) of human so ial networks. We ompare generated networks with a large-s ale so ial network data set, validating that the model is able to produ e graphs with the same stru tural proper- ties of human-so ial-network graphs. Moreover, we hara terise the im- pa t of the model parameters on the syntheti graph properties.