Link Creation and Profile Alignment in the aNobii Social Network

Abstract

Abstract-The present work investigates the structural and dynamical properties of aNobii 1 , a social bookmarking system designed for readers and book lovers. Users of aNobii provide information about their library, reading interests and geographical location, and they can establish typed social links to other users. Here, we perform an in-depth analysis of the system's social network and its interplay with users' profiles. We describe the relation of geographic and interest-based factors to social linking. Furthermore, we perform a longitudinal analysis to investigate the interplay of profile similarity and link creation in the social network, with a focus on triangle closure. We report a reciprocal causal connection: profile similarity of users drives the subsequent closure in the social network and, reciprocally, closure in the social network induces subsequent profile alignment. Access to the dynamics of the social network also allows us to measure quantitative indicators of preferential linking.

Temporal Patterns of Communication in Social Networks

“…The first category is composed by vertex attributes, which generally includes nodes' (in and out) degrees, overall aggregated activity (or strength), usually measured as the total number of communication events and centrality measures as the clustering coefficient. In general, vertex features are positively correlated with the creation of a tie and contribute significantly to the predictive power of the model, a result that reflects assortative mixing and correlations in the degree and the activity discussed in the previous chapters (Aiello et al 2010;Liben-Nowell and Kleinberg 2007;Raeder et al 2011). Predictive models significantly improve when taking into account also dyad-level attributes (or tie-level attributes) (Lü and Zhou 2010;Raeder et al 2011;Zhou et al 2009).…”

“…The most exemplary work on edge decay in social networks is probably that of Burt who studied the social networks of the most important bankers over time and analyzed those factors that contribute to the disappearance of edges between them (Burt 2000;Burt 2002) . Several other studies have focused on link characterization in other communication networks as phone and SMS networks (Akoglu and Dalvi 2010;Hidalgo and Rodriguez-Sickert 2008;Raeder et al 2011), online networks (Aiello et al 2010;Crandall et al 2008;Gilbert and Karahalios 2009;Kivran-Swaine et al 2011;Romero et al 2011) and off-line settings (Burt 2000;Martin and Yeung 2006). One of the main substantive conclusions that emerge from these studies is that links exhibit a memory, meaning that old links are more likely to persist in time than newlyformed ones (where the "age" and persistence of a link is defined in terms of observed communication events or when the link is registered or deleted in on-line social networks).…”

“…The conventional way of addressing this problem is based on the assumption that many real networks are endogenous, thus their properties at a given time contain information on the status of the networks in the future. Although a number of factors, both endogenous and exogenous usually contribute towards tie creation and decay in social networks, there is a vast literature suggesting that information about future interactions can be actually extracted from the network structure (Aiello et al 2010;Akoglu and Dalvi 2010;Eagle et al 2009;Hidalgo and Rodriguez-Sickert 2008;Liben-Nowell and Kleinberg 2007;Kossinets and Watts 2006;Raeder et al 2011).…”

Community detection in networks: Structural communities versus ground truth

“…The Amazon product copurchasing network (amazon) has groups of product categories [14]. anobii is a book recommendation social network popular in Italy, where users can join groups [15,16]. The dblp network of coauthorships in computer science literature has publication venues as groups [17].…”

Combining qualitative, quantitative and structural dimensions in a longitudinal perspective. The case of network influence

“…Acknowledgement of influence is, after all, problematic. Identification of the enabling vectors and their dissociation from selection processes is an important task for sociologists and other scientists (Crandall, Cosley and al., 2008;Aiello, Barrat and al, 2010;Steglich, Snijders, Pearson, 2010;Mercken, Snijders and al., 2010). Consequently, we examine here the issue from various perspectives.…”