Citation Context

...actor. In this paper, we adopt an egocentric approach, in which the local network of a central ego is expressed as a set of alter nodes connected to the ego, together with all the links among alters (=-=Wasserman and Faust 1994-=-). This incorporates a full set of ego’s acquaintances, allowing us to estimate the extent to which the friend set is segmented into distinct groups. According to homophily, similar people are more li...

Citation Context

...ion of u’s neighbors; they are equivalent when Cku contains all u’s friends. Table 1: Densities of clusters on various topics averaged across all sampled ego networks. We use the Mann-Whitney U test (=-=Mann and Whitney 1947-=-) to check whether differences between empirical measures and the two baselines are statistically significant. 〈D〉u is compared with both 〈D1〉u and 〈D2〉u, and the bigger p-value is displayed. The larg...

Citation Context

...(www.aaai.org). All rights reserved. Related Work Most existing research into homophily focuses on dyadic measures, pre-defined social groups, or global patterns in the network (Crandall et al. 2008; =-=Backstrom et al. 2006-=-; Schifanella et al. 2010; Aiello et al. 2012). While these approaches provide a great deal of information about homophily, influence, and similarity, they do not provide an analysis from the perspect...

Citation Context

...McPherson, Lovin, and Cook 2001; Kossinets and Watts 2009). When combined with social influence processes, this can result in a feedback loop that produces increasingly homogeneous personal networks (=-=Crandall et al. 2008-=-; Jamieson and Cappella 2009). Indeed, theoretical models of social influence propose that extreme polarization of interests can evolve even within a population of actors who hold diverse sets of opin...

Citation Context

...th rates, and little overlap. Introduction The principle of homophily states that people with common characteristics are more likely to have contact with one another (McPherson, Lovin, and Cook 2001; =-=Kossinets and Watts 2009-=-). When combined with social influence processes, this can result in a feedback loop that produces increasingly homogeneous personal networks (Crandall et al. 2008; Jamieson and Cappella 2009). Indeed...

Citation Context

...hts reserved. Related Work Most existing research into homophily focuses on dyadic measures, pre-defined social groups, or global patterns in the network (Crandall et al. 2008; Backstrom et al. 2006; =-=Schifanella et al. 2010-=-; Aiello et al. 2012). While these approaches provide a great deal of information about homophily, influence, and similarity, they do not provide an analysis from the perspective of a single actor. In...

Citation Context

...y 2010; Romero, Tan, and Ugander 2013; Weng and Menczer 2014). Topical locality on the Web describes a phenomenon similar to homophily: most Web pages tend to link with related content (Davison 2000; =-=Menczer 2004-=-). This concept can be extended to interpersonal interaction. For example, social bookmarking services provide a venue for collaborative tagging behavior, producing an emergent social network built up...

Citation Context

...to explain an increase in homogeneity within networks (Crandall et al. 2008; Jamieson and Cappella 2009). The existence of homophily in online settings has been observed in various empirical studies (=-=Fiore and Donath 2005-=-; Aral, Muchnik, and Sundararajan 2009; De Choudhury 2011), although dissimilarity, disagreement, and heterogeneity can also exist among people close to each other, yielding division among social grou...

Citation Context

...eson and Cappella 2009). Indeed, theoretical models of social influence propose that extreme polarization of interests can evolve even within a population of actors who hold diverse sets of opinions (=-=Macy et al. 2003-=-; Flache and Macy 2011). Meanwhile, social circles tend to restrict the range of interests that two people may discuss, further enhancing the potential polarization of interest groups in ego networks;...

Citation Context

...Cook 2001; Kossinets and Watts 2009). When combined with social influence processes, this can result in a feedback loop that produces increasingly homogeneous personal networks (Crandall et al. 2008; =-=Jamieson and Cappella 2009-=-). Indeed, theoretical models of social influence propose that extreme polarization of interests can evolve even within a population of actors who hold diverse sets of opinions (Macy et al. 2003; Flac...

Citation Context

...k Most existing research into homophily focuses on dyadic measures, pre-defined social groups, or global patterns in the network (Crandall et al. 2008; Backstrom et al. 2006; Schifanella et al. 2010; =-=Aiello et al. 2012-=-). While these approaches provide a great deal of information about homophily, influence, and similarity, they do not provide an analysis from the perspective of a single actor. In this paper, we adop...

Citation Context

...during interpersonal conversation. Topics and user interests haven been studied broadly in online environments from multiple perspectives (Michelson and Macskassy 2010; Romero, Tan, and Ugander 2013; =-=Weng and Menczer 2014-=-). Topical locality on the Web describes a phenomenon similar to homophily: most Web pages tend to link with related content (Davison 2000; Menczer 2004). This concept can be extended to interpersonal...

Citation Context

...2009). Indeed, theoretical models of social influence propose that extreme polarization of interests can evolve even within a population of actors who hold diverse sets of opinions (Macy et al. 2003; =-=Flache and Macy 2011-=-). Meanwhile, social circles tend to restrict the range of interests that two people may discuss, further enhancing the potential polarization of interest groups in ego networks; for example, family m...