The strength of co-authorship ties through different topological properties

Abstract

Social networks are complex structures that describe individuals (graph nodes) connected in any social context (graph edges). Different metrics can be applied to those networks and their properties in order to understand behavior and even predict the future. One of such properties is tie strength, which allows to identify prominent individuals, analyze how relationships play different roles, predict links, and so on. Here, we specifically address the problem of measuring tie strength in co-authorship social networks (nodes are researchers and edges represent their co-authored publications). We start by presenting four cases that emphasize the problems of current metrics. Then, we propose a new metric for tie strength, called tieness, that is simple to calculate and better differentiates the degrees of strength. Accompanied with a nominal scale, tieness also provides better results when compared to the existing metrics. Our analyses consider three real social networks built from publications collected from digital libraries on Computer Science, Medicine, and Physics. Finally, we also make all datasets publicly available.

Who is really in my social circle?

“…In practice, simple metrics based on these ideas, such as frequency of interaction [8] and number of mutual friends [10], are commonly used to measure tie strength. Nevertheless, depending on the context, other factors can be considered when modelling tie strength [11][12][13][14][15]. Adamic et al [14] consider information such as membership of mailing lists and use of common phrases on personal Web pages to measure the similarity between users (or tie strength).…”

“…Adamic et al [14] consider information such as membership of mailing lists and use of common phrases on personal Web pages to measure the similarity between users (or tie strength). Brandão et al [15] measure the strength of coauthorship ties based on topological properties of academic networks. Gilbert and Karahalios [11] propose a tie strength metric based on specific information extracted from Facebook, such as the number of days since the last communication and the number of words exchanged in Wall posts.…”

“…Thus, for a given value of κ, V κ includes all vertices that interacted in the κth period of time. The edges in the set E κ represent the interactions between pairs of vertex (v i , v j ) that occurred during the time period κ Here, we apply this temporal graph model to real social networks of scientific collaborations in the period 2000-2016, built from APS, DBLP and PubMed datasets 4 [15], university campus mobility networks from Dartmouth College and USC [17], and an e-mail network derived from communications among Enron's employees from 1999 to 2003 5 . Table 1 presents a general characterization of these networks.…”

“…4 In this context, a module is a community that is characterized by high modularity. 9 Real networks are usually sparse with the number of edges increasing linearly with the number of vertices [80] Availability of data and materials The datasets used and analyzed during the current study have also been used elsewhere: [15] (APS, PubMed, DBLP), [72] (Dartmouth), [73] (USC) and [30,74] (Enron). The generated datasets are available from the corresponding author on reasonable request in the http://cnet.jcloud.net.br repository.…”

Universidade Federal de Minas Gerais - Belo Horizonte - MG

“…co-authorship frequency or edge weight) have been largely used to measure the strength of ties. However, through empirical analyses, we identified four main problems with using solely neighborhood overlap and co-authorship frequency to measure tie strength [Brandão et al 2016, Brandão andMoro 2017b]: (Case 1) when a pair of collaborators does not have any common neighbor, neighborhood overlap will be zero; (Case 2) when determining if two collaborators are from the same community or not, co-authorship frequency considers only the absolute frequency of interaction; (Case 3) when there is little collaboration between a pair of collaborators and manyn common neighbors, neighborhood overlap and co-authorship frequency will present opposite results; and (Case 4) when the results are extreme values, neighborhood overlap may not represent the reality.…”