Search

This paper proposes a Bayesian multifactor spatio-temporal model with a dynamic spatial autoregressive coefficient to estimate time-varying network interdependence with time-series cross-sectional data. To correct biases caused by unobserved time-varying homophily and common shocks with heterogeneous impact, we adopt the multifactor method to control for latent confounders and tackle the classic ``reflection problem." We use Bayesian shrinkage techniques to select latent factors. Monte Carlo studies and two empirical examples demonstrate that the multifactor specification reduces biases and improves the identification of network interdependence, and that falsely assuming a time-invariant network effect not only misses the variation of interdependence but also mis-estimates its central tendency. The proposed model is robust to misspecifcation and can accommodate various types of social networks.

This article examines peer influences from network relationships within a social network game (i.e., embeddedness) and across such games (i.e., multiplexity). Drawing on social influence theory, we develop a bivariate Poisson model of users’ repeated visits and latent attrition that accommodates peer interaction after controlling for homophily. We estimate the model using data from two social network games with considerable overlap among network members. We find that friends who are only multiplex across games exert greater peer influence on users’ game visits than members who are embedded within a single game. We also determined that ignoring network multiplexity across games may lead firms to mistarget users due to biased peer influences of embedded friends. This result provides an unresearched explanation—strength of peer influence—for the mixed findings in previous literature on network embeddedness. We utilized our results to conduct several scenario analyses to demonstrate how firms can effectively manage users’ engagement and target users in multiple social network games.

Understanding actor collaboration networks and their evolution is essential to promoting collective action in resilience planning and management of interdependent infrastructure systems. Local interactions and choice homophily are two important network evolution mechanisms. Network motifs encode the information of network formation, configuration, and the local structure. Homophily effects, on the other hand, capture whether the network configurations have significant correlations with node properties. The objective of this paper is to explore the extent to which local interactions and homophily effects influence actor collaboration in resilience planning and management of interdependent infrastructure systems. We mapped bipartite actor collaboration network based on a post-Hurricane Harvey stakeholder survey that revealed actor collaborations for hazard mitigation. We examined seven bipartite network motifs for the mapped collaboration network and compared the mapped network to simulated random models with same degree distributions. Then we examined whether the network configurations had significant statistics for node properties using exponential random graph models. The results provide insights about the two mechanisms—local interactions and homophily effect—influencing the formation of actor collaboration in resilience planning and management of interdependent urban systems. The findings have implications for improving network cohesion and actor collaborations from diverse urban sectors.

Graph Neural Networks (GNNs) are effective in many applications. Still, there is a limited understanding of the effect of common graph structures on the learning process of GNNs. To fill this gap, we study the impact of community structure and homophily on the performance of GNNs in semi-supervised node classification on graphs. Our methodology consists of systematically manipulating the structure of eight datasets, and measuring the performance of GNNs on the original graphs and the change in performance in the presence and the absence of community structure and/or homophily. Our results show the major impact of both homophily and communities on the classification accuracy of GNNs, and provide insights on their interplay. In particular, by analyzing community structure and its correlation with node labels, we are able to make informed predictions on the suitability of GNNs for classification on a given graph. Using an information-theoretic metric for community-label correlation, we devise a guideline for model selection based on graph structure. With our work, we provide insights on the abilities of GNNs and the impact of common network phenomena on their performance. Our work improves model selection for node classification in semi-supervised settings.

Trust prediction provides valuable support for decision making, information dissemination, and product promotion in online social networks. As a complex concept in the social network community, trust relationships among people can be established virtually based on: 1) their interaction behaviors, e.g., the ratings and comments that they provided; 2) the contextual information associated with their interactions, e.g., location and culture; and 3) the relative temporal features of interactions and the time periods when the trust relationships hold. Most of the existing works only focus on some aspects of trust, and there is not a comprehensive study of user trust development that considers and incorporates 1)-3) in trust prediction. In this article, we propose a context-aware deep trust prediction model C-DeepTrust to fill this gap. First, we conduct user feature modeling to obtain the user's static and dynamic preference features in each context. Static user preference features are obtained from all the ratings and reviews that a user provided, while dynamic user preference features are obtained from the items rated/reviewed by the user in time series. The obtained context-aware user features are then combined and fed into the multilayer projection structure to further mine the context-aware latent features. Finally, the context-aware trust relationships between users are calculated by their context-aware feature vector cosine similarities according to the social homophily theory, which shows a pervasive property of social networks that trust relationships are more likely to be developed among similar people. Extensive experiments conducted on two real-world datasets show the superior performance of our approach compared with the representative baseline methods.

Social influence cannot be identified from purely observational data on social networks, because such influence is generically confounded with latent homophily, i.e., with a node's network partners being informative about the node's attributes and therefore its behavior. If the network grows according to either a latent community (stochastic block) model, or a continuous latent space model, then latent homophilous attributes can be consistently estimated from the global pattern of social ties. We show that, for common versions of those two network models, these estimates are so informative that controlling for estimated attributes allows for asymptotically unbiased and consistent estimation of social-influence effects in linear models. In particular, the bias shrinks at a rate which directly reflects how much information the network provides about the latent attributes. These are the first results on the consistent non-experimental estimation of social-influence effects in the presence of latent homophily, and we discuss the prospects for generalizing them., Comment: 35 pages, 4 figures

Predictions of node categories are commonly used to estimate homophily and other relational properties in networks. However, little is known about the validity of using predictions for this task. We show that estimating homophily in a network is a problem of predicting categories of dyads (edges) in the graph. Homophily estimates are unbiased when predictions of dyad categories are unbiased. Node-level prediction models, such as the use of names to classify ethnicity or gender, do not generally produce unbiased predictions of dyad categories and therefore produce biased homophily estimates. Bias comes from three sources: sampling bias, correlation between model errors and node degree, and correlation between node-level model errors along dyads. We examine three methods for estimating homophily: predicting node categories, predicting dyad categories, and a hybrid “ego–alter” approach. This analysis indicates that only the dyadic prediction approach is unbiased, whereas the node-level approach produces both high bias and high overall error. We find that node-level classification performance is not a reliable indicator of accuracy for homophily. Although this article focuses on a particular version of homophily, results generalize to heterophilous cases and other dyadic measures. We conclude with suggestions for research design. Code for this article is available at https://github.com/georgeberry/autocorr.

Social networks display homophily and clustering, and are usually sparse. I develop and estimate a structural model of strategic network formation with heterogeneous players and latent community structure, whose equilibrium networks are sparse and exhibit homophily and clustering. Each player belongs to a community unobserved by the econometrician. Players' payoffs vary by community and depend on the composition of direct links and common neighbors, allowing preferences to have a bias for similar people. Players meet sequentially and decide whether to form bilateral links, after receiving a random matching shock. The probability of meeting people in different communities is smaller than the probability of meeting people in the same community, and it decreases with the size of the network. The model converges to an exponential family random graph, with weak dependence among links. As a consequence the equilibrium networks are sparse and the sufficient statistics of the network are asymptotically normal. The posterior distribution of structural parameters and unobserved heterogeneity is estimated with school friendship network data from Add Health, using a Bayesian exchange algorithm. The estimates detect high levels of racial homophily, and heterogeneity in both costs of links and payoffs from common friends. The posterior predictions show that the model is able to replicate the homophily levels and the aggregate clustering of the observed network, in contrast with standard exponential family network models.

Social Internet of Things (SIoT) is a variation of social networks that adopt the property of peer-to-peer networks, in which connections between the things and social actors are automatically established. SIoT is a part of various organizations that inherit the social interaction, and these organizations include industries, institutions, and other establishments. Triadic closure and homophily are the most commonly used measures to investigate social networks’ formation and nature, where both measures are used exclusively or with statistical models. The triadic closure patterns are mapped for actors’ communication behavior over a location-based social network, affecting the homophily. In this study, we investigate triads emergence in homophilic social networks. This evaluation is based on the empirical review of triads within social networks (SNs) formed on Big Data. We utilized a large location-based dataset for an in-depth analysis, the Chinese telecommunication-based anonymized call detail records (CDRs). Two other openly available datasets, Brightkite and Gowalla, were also studied. We identified and proposed three social triad classes in a homophilic network to feature the correlation between social triads and homophily. The study opened a promising research direction that relates the variation of homophily based on closure triads nature. The homophilic triads are further categorized into transitive and intransitive groups. As our concluding research objective, we examined the relative triadic throughput within a location-based social network for the given datasets. The research study attains significant results highlighting the positive connection between homophily and a specific social triad class.

Recommender systems can benefit from a plethora of signals influencing user behavior such as her past interactions, her social connections, as well as the similarity between different items. However, existing methods are challenged when taking all this data into account and often do not exploit all available information. This is primarily due to the fact that it is non-trivial to combine the various information as they mutually influence each other. To address this shortcoming, here, we propose a ‘Fusion Recommender’ (FuseRec), which models each of these factors separately and later combines them in an interpretable manner. We find this general framework to yield compelling results on all three investigated datasets, Epinions, Ciao, and CiaoDVD, outperforming the state-of-the-art by more than 14% for Ciao and Epinions. In addition, we provide a detailed ablation study, showing that our combined model achieves accurate results, often better than any of its components individually. Our model also provides insights on the importance of each of the factors in different datasets.

Unveil the homophilic/heterophilic behaviors that characterize the wiring patterns of complex networks is an important task in social network analysis, often approached studying the assortative mixing of node attributes. Recent works underlined that a global measure to quantify node homophily necessarily provides a partial, often deceiving, picture of the reality. Moving from such literature, in this work, we propose a novel measure, namely Conformity, designed to overcome such limitation by providing a node-centric quantification of assortative mixing patterns. Differently from the measures proposed so far, Conformity is designed to be path-aware, thus allowing for a more detailed evaluation of the impact that nodes at different degrees of separations have on the homophilic embeddedness of a target. Experimental analysis on synthetic and real data allowed us to observe that Conformity can unveil valuable insights from node-attributed graphs., Comment: Submitted to IEEE Intelligent Systems

Graph neural networks (GNNs) have been the dominant approaches for graph representation learning. However, most GNNs are applied to homophily graphs and perform poorly on heterophily graphs. Meanwhile, these GNNs fail to directly capture long-range dependencies and complex interactions between 1-hop neighbors when generating node representations by iteratively aggregating directly connected neighbors. In addition, structural patterns, such as motifs which have been established as building blocks for graph structure, contain rich topological and semantical information and are worth studying further. In this paper, we introduce the common-neighbors based motifs, which we called CN-motifs, to generalize and enrich the definition of structural patterns. We group the 1-hop neighbors and construct a high-order graph according to CN-motifs, and propose CN-motifs Perceptive Graph Neural Networks (CNMPGNN), a novel framework which can effectively resolve problems mentioned above. Notably, by making full use of structural patterns, our model achieves the state-of-the-art results on several homophily and heterophily datasets.

The networking among organizations and within the organization is very important to achieve vision, mission and the performance of any organization. Although there are studies about networking in a global context, it is hard to find networking related studies in the local context. Therefore, this study of networking within divisional secretaries and its impact on their performance is important. The main objective of this study is to analysis the networking within the divisional secretaries and its impact on the performance of those divisional secretaries. The studies depend upon the answers based on the questionnaire survey. Although the networks fall into two principal categories such as inter-firm networks and inter-personal networks, this study is limited to studying interpersonal network. The network is a vague concept and it consists of a set of actors (“nodes”) and the relations (“ties” or “edges”) between these actors. Relevant literature discussed to build the analytical framework and different theories of networking such as theories of self-interest, social exchange and dependency theory, mutual interest and collective action, a family of cognitive theories and theory of homophily. It reviewed the relevant literature on job performance and networking related performance. It also discussed the divisional administration and divisional secretaries. In this study impact of networking on the performance of divisional secretaries was measured with five independent variables such as instrumental network, expressive network, friendship network, advice network and knowledge network. Primary and secondary data were used to analyze the networking related performance. The survey was done throughout the all divisional secretaries in Sri Lanka via a questionnaire and collected data were analyzed by using Structural equation modelling (SEM). According to the study, the most important and affected networking activity which was highly affected by the performance is the knowledge network. Friendship network and instrumental network are the second and third networks effects on performance that is to follow on respectively. The expressive network has a moderate effect on performance while the advice network shows no positive effect on performance.

The validity of network observations is sometimes of concern in empirical studies, since observed networks are prone to error and may not represent the population of interest. This lack of validity is not just a result of random measurement error, but often due to systematic bias that can lead to the misinterpretation of actors’ preferences of network selections. These issues in network observations could bias the estimation of common network models (such as those pertaining to influence and selection) and lead to erroneous statistical inferences. In this study, we proposed a simulation-based sensitivity analysis method that can evaluate the robustness of inferences made in social network analysis to six forms of selection mechanisms that can cause biases in network observations—random, homophily, anti-homophily, transitivity, reciprocity, and preferential attachment. We then applied this sensitivity analysis to test the robustness of inferences for social influence effects, and we derived two sets of analytical solutions that can account for biases in network observations due to random, homophily, and anti-homophily selection.

In real world, social networks are often homophily-driven evolving, which represents the situation that users cut old connections and try to connect with others who share same attributes with them. However, existing works about information diffusion mainly focus on the static social network, while the influences of homophily-driven evolution has been neglected. Motivated by this, we investigate the diffusion accuracy problem in homophily-driven evolving social networks. Specifically, we consider a spreading-based diffusion mechanism, where a user simply spreads the information she/he is interested in to all her/his friends. This spreading-based diffusion mechanism is blind-guided and results in low diffusion performance in social networks without homophily-driven evolution. Our theoretical analyses present that the diffusion accuracy can be greatly improved during the evolution process. Moreover, we disclose that when the evolution process converges to a stable state, the diffusion process could achieve even higher performance, where all the information receivers are interested in it. In other word, the diffusion accuracy can simultaneously achieve high precision and recall. At last, the theoretical results are verified by simulations based on the synthetic network and experimental results based on real world network.

The similarity between pairs of people is often measured on relatively static traits and at a given point in time. Moving beyond this approach, a burgeoning line of research is investigating temporal dyadic similarity on traits and behaviors, such as health activities. Our study contributes to this line of inquiry by using fine-grained longitudinal data obtained from sensors, mobile devices, and surveys to examine whether we can observe distinct types of dyadic similarity trajectories based on physical activity, and if so, what dyad-level properties predict membership in each trajectory class. Treating daily differences in the steps for dyads as time series, we use k-shape clustering to identify classes of similarity trajectories and logistic regression to examine the link between trajectory class and key dyad-level factors. We identify 21 dyadic trajectory clusters and find that trajectory membership predicts dyadic connectivity in the communication network, as well as racial and religious—but not gender-based—similarity. We conclude by noting how research on dyadic similarity trajectories can be further integrated with ongoing work in social network analysis.

Fragmentation, echo chambers, and their amelioration in social networks have been a growing concern in the academic and non-academic world. This paper shows how, under the assumption of homophily, echo chambers and fragmentation are system-immanent phenomena of highly flexible social networks, even under ideal conditions for heterogeneity. We achieve this by finding an analytical, network-based solution to the Schelling model and by proving that weak ties do not hinder the process. Furthermore, we derive that no level of positive algorithmic bias in the form of rewiring is capable of preventing fragmentation and its effect on reducing the fragmentation speed is negligible., Cite as: Chris Blex & Taha Yasseri (2020) Positive algorithmic bias cannot stop fragmentation in homophilic networks, The Journal of Mathematical Sociology, DOI: 10.1080/0022250X.2020.1818078

As a fundamental problem in social network analysis, community detection has recently attracted wide attention, accompanied by the output of numerous community detection methods. However, most existing methods are developed by only exploiting link topology, without taking node homophily (i.e., node similarity) into consideration. Thus, much useful information that can be utilized to improve the quality of detected communities is ignored. To overcome this limitation, we propose a new community detection approach based on nonnegative matrix factorization (NMF), namely, homophily preserving NMF (HPNMF), which models not only link topology but also node homophily of networks. As such, HPNMF is able to better reflect the inherent properties of community structure. In order to capture node homophily from scratch, we provide three similarity measurements that naturally reveal the association relationships between nodes. We further present an efficient learning algorithm with convergence guarantee to solve the proposed model. Finally, extensive experiments are conducted, and the results demonstrate that HPNMF has strong ability to outperform the state-of-the-art baseline methods.

During the formation process of stochastic networks, nodes tend to establish edges based on selective linkage mechanisms. In general, these mechanisms involve probability distributions that underlie the selection of target nodes. In social networks, edges are often associated to relationships that are homophilic with respect to individual traits. Such traits include, for example, gender and age, and are referred to as node types. Our work considers an affinity-weighted preferential attachment model that characterizes the tendency of two group of nodes to connect to other nodes of the same type. We derive mathematical expressions for the dynamics and convergence of homophily measures at the node, group, and network level. Furthermore, we characterize the convergence of network modularity and show that the formation of community structures can be expressed as a function of network homophily.

In network models for real-world domains often network adaptation has to be addressed by incorporating certain network adaptation principles. In some cases, also higher-order adaptation occurs: the adaptation principles themselves also change over time. To model such multilevel adaptation processes it is useful to have some generic architecture. Such an architecture should describe and distinguish the dynamics within the network (base level), but also the dynamics of the network itself by certain adaptation principles (first-order adaptation level), and also the adaptation of these adaptation principles (second-order adaptation level), and maybe still more levels of higher-order adaptation. This paper introduces a multilevel network architecture for this, based on the notion network reification. Reification of a network occurs when a base network is extended by adding explicit states representing the characteristics of the structure of the base network. It will be shown how this construction can be used to explitly represent network adaptation principles within a network. When the reified network is itself also reified, also second-order adaptation principles can be explicitly represented. The multilevel network reification construction introduced here is illustrated for an adaptive adaptation principle from Social Science for bonding based on homophily. This first-order adaptation principle describes how connections are changing, whereas this first-order adaptation principle itself changes over time by a second-order adaptation principle.

Relationship management has been of strategic importance for businesses that are interested to evaluate the state of the relationship with the customer and if possible to migrate customers to better and more binding states. This work addresses the problem of estimating the relationship state of a customer and examining the migration policy of the customer, using social media analytics. We propose an innovative framework, where clustering, linguistic and emotional analytics are used to automatically assign users to relationship states. Our research is of multi-disciplinary nature, where we are using existing results from surveys on users’ behavior when mitigating states to verify the semantics of our metrics, showing that they follow similar behavior. Our results show that clustering users based on communication, emotions and perceived product mix can result in an automated assignment of users to states. Furthermore, trust, commitment and homophily are defined and our results show that users are migrating states influenced by these values. Our work provides data analytics metrics for businesses that will identify and address the problem of relationship management thus improving the overall users’ satisfaction using a data analytics approach.

Bringing together considerations from three research trends (honest signals of collaboration, socio-semantic networks and homophily theory), we hypothesise that word use similarity and having similar social network positions are linked with the level of employees’ digital interaction. To verify our hypothesis, we analyse the communication of close to 1600 employees, interacting on the intranet communication forum of a large company. We study their social dynamics and the ‘honest signals’ that, in past research, proved to be conducive to employees’ engagement and collaboration. We find that word use similarity is the main driver of interaction, much more than other language characteristics or similarity in network position. Our results suggest carefully choosing the language according to the target audience and have practical implications for both company managers and online community administrators. Understanding how to better use language could, for example, support the development of knowledge sharing practices or internal communication campaigns.

We investigate how individuals form social connections to access information. In our link formation game of one-way flows, the myopic best response is to link to the node with the highest informational reach, which depends on indirect links. This strategy leads to the efficient circle network. We find that myopic best response predicts the links that subjects form. However, some subjects target high-degree nodes instead, neglecting indirect connections. This reduces network efficiency. We obtain similar findings when subjects link to transfer information to others. Using a minimal group treatment, we also find evidence of in-group homophily in link formation.

This study formulates a model where (i) players are characterized by a knowledge set that changes endogenously by communication and (ii) some players have homophily preferences, while others have h...

It is widely observed that individuals prefer to interact with others who are more similar to them (this phenomenon is termed homophily). This similarity manifests itself in various ways such as beliefs, values and education. Thus, it should not come as a surprise that when people make hiring choices, for example, their similarity to the candidate plays a role in their choice. In this paper, we suggest that putting the decision in the hands of a committee instead of a single person can reduce this bias. We study a novel model of voting in which a committee of experts is constructed to reduce the biases of its members. We first present voting rules that optimally reduce the biases of a given committee. Our main results include the design of committees, for several settings, that are able to reach a nearly optimal (unbiased) choice. We also provide a thorough analysis of the trade-offs between the committee size and the obtained error. Our model is inherently different from the well-studied models of voting that focus on aggregation of preferences or on aggregation of information due to the introduction of similarity biases.

In recent years, with the growing number of online social networks, these networks have become one of the best markets for advertising and commerce, so studying these networks is very important. Forecasting new edges in online social networks can give us a better understanding of the growth of these networks. There have been many studies of link prediction in the field of engineering and humanities. Scientists attribute the existence of a new relationship between two individuals for two reasons: 1) Proximity to the graph (structure) 2) Similar properties of the two individuals (Homophile law). However, studying the impact of the two approaches working together to create new edges remains an open problem. Similarity metrics can also be divided into two categories; Neighborhood-based and path-based. So far, the above two theoretical approaches (proximity and homophile) have not been found together in the neighborhood-based metrics. In this paper, we first attempt to provide a solution to determine importance of the proximity to the graph and similar features in the connectivity of the graphs. Then obtained weights are assigned to both proximity and homophile. Then the best similarity metric in each approach are obtained. Finally, the selected metric of homophily similarity and structural similarity are combined with the obtained weights. The results of this study were evaluated on two datasets; Zanjan University Graduate School of Social Sciences and Pokec online Social Network. The first data set was collected for this study and then the questionnaires and data collection methods were filled out. Since this dataset is one of the few Iranian datasets that has been compiled with its users' specifications, it can be of great value. In this paper, we have been able to increase the accuracy of Neighborhood-based similarity metric by using two proximity in graph and homophily approaches., Comment: in Farsi. Quarterly Journal Signal and Data Processing (JSDP), 2020

This chapter covers the second part of our business intelligence discussion and makes the reader learn how organizations can create business value by analyzing social network data. Diverse information about a certain person can be collected from different social media tools and combined into a database to obtain more complete profiles of employees, customers, or prospects (i.e., social engineering). The latter can supplement the social CRM database (see Chap. 5). Particularly, social media may uncover information about what people post, share, or like but also to whom they are connected. By combining or aggregating such information for many individuals in social networks, organizations can start predicting trends, e.g., to improve their targeted marketing (see Chap. 4) or to predict which people are more likely to churn, fraud, resign, etc. Hence, social media are seen as big data in the sense that they can provide massive amounts of real-time data about many Internet users, which can be used to predict someone’s future behavior based on past behavior of others. This chapter explains how social networks can be built from social media data and introduces concepts such as peer influence and homophily. The chapter concludes with big data challenges to social network data.

In sociological research, the study of macro processes, such as opinion polarization, faces a fundamental problem, the so-called micro-macro problem. To overcome this problem, we combine empirical experimental research on biased argument processing with a computational theory of group deliberation in order to clarify the role of biased processing in debates around energy. The experiment reveals a strong tendency to consider arguments aligned with the current attitude more persuasive and to downgrade those speaking against it. This is integrated into the framework of argument communication theory in which agents exchange arguments about a certain topic and adapt opinions accordingly. We derive a mathematical model that allows to relate the strength of biased processing to expected attitude changes given the specific experimental conditions and find a clear signature of moderate biased processing. We further show that this model fits significantly better to the experimentally observed attitude changes than the neutral argument processing assumption made in previous models. Our approach provides new insight into the relationship between biased processing and opinion polarization. At the individual level our analysis reveals a sharp qualitative transition from attitude moderation to polarization. At the collective level we find (i.) that weak biased processing significantly accelerates group decision processes whereas (ii.) strong biased processing leads to a persistent conflictual state of subgroup polarization. While this shows that biased processing alone is sufficient for the emergence of polarization, we also demonstrate that homophily may lead to intra-group conflict at significantly lower rates of biased processing.

Applied network science 6(1), 78 (2021). doi:10.1007/s41109-021-00394-3, Published by Springer International Publishing, [Cham]

Graph convolutional networks (GCNs) have achieved great success in many applications and have caught significant attention in both academic and industrial domains. However, repeatedly employing graph convolutional layers would render the node embeddings indistinguishable. For the sake of avoiding oversmoothing, most GCN-based models are restricted in a shallow architecture. Therefore, the expressive power of these models is insufficient since they ignore information beyond local neighborhoods. Furthermore, existing methods either do not consider the semantics from high-order local structures or neglect the node homophily (i.e., node similarity), which severely limits the performance of the model. In this article, we take above problems into consideration and propose a novel Semantics and Homophily preserving Network Embedding (SHNE) model. In particular, SHNE leverages higher order connectivity patterns to capture structural semantics. To exploit node homophily, SHNE utilizes both structural and feature similarity to discover potential correlated neighbors for each node from the whole graph; thus, distant but informative nodes can also contribute to the model. Moreover, with the proposed dual-attention mechanisms, SHNE learns comprehensive embeddings with additional information from various semantic spaces. Furthermore, we also design a semantic regularizer to improve the quality of the combined representation. Extensive experiments demonstrate that SHNE outperforms state-of-the-art methods on benchmark datasets.

Recommender systems serve as mediators of information consumption and propagation. In this role, these systems have been recently criticized for introducing biases and promoting the creation of echo chambers and filter bubbles, thus lowering the diversity of both content and potential new social relations users are exposed to. Some of these issues are a consequence of the fundamental concepts on which recommender systems are based on. Assumptions like the homophily principle might lead users to content that they already like or friends they already know, which can be naive in the era of ideological uniformity and fake news. A significant challenge in this context is how to effectively learn the dynamic representations of users based on the content they share and their echo chamber or community interactions to recommend potentially relevant and diverse friends from outside the network of influence of the users’ echo chamber. To address this, we devise FRediECH (a Friend RecommenDer for breakIng Echo CHambers), an echo chamber-aware friend recommendation approach that learns users and echo chamber representations from the shared content and past users’ and communities’ interactions. Comprehensive evaluations over Twitter data showed that our approach achieved better performance (in terms of relevance and novelty) than state-of-the-art alternatives, validating its effectiveness.

In order to navigate a complex web of relationships, an individual must learn and represent the connections between people in a social network. However, the sheer size and complexity of the social world makes it impossible to acquire firsthand knowledge of all relations within a network, suggesting that people must make inferences about unobserved relationships to fill in the gaps. Across three studies (n = 328), we show that people can encode information about social features (e.g., hobbies, clubs) and subsequently deploy this knowledge to infer the existence of unobserved friendships in the network. Using computational models, we test various feature-based mechanisms that could support such inferences. We find that people’s ability to successfully generalize depends on two representational strategies: a simple but inflexible similarity heuristic that leverages homophily, and a complex but flexible cognitive map that encodes the statistical relationships between social features and friendships. Together, our studies reveal that people can build cognitive maps encoding arbitrary patterns of latent relations in many abstract feature spaces, allowing social networks to be represented in a flexible format. Moreover, these findings shed light on open questions across disciplines about how people learn and represent social networks and may have implications for generating more human-like link prediction in machine learning algorithms.

Understanding why people join, stay, or leave social groups is a central question in the social sciences, including computational social systems, while modeling these processes is a challenge in complex networks. Yet, the current empirical studies rarely focus on group dynamics for lack of data relating opinions to group membership. In the NetSense data, we find hundreds of face-to-face groups whose members make thousands of changes of memberships and opinions. We also observe two trends: opinion homogeneity grows over time, and individuals holding unpopular opinions frequently change groups. These observations and data provide us with the basis on which we model the underlying dynamics of human behavior. We formally define the utility that members gain from ingroup interactions as a function of the levels of homophily of opinions of group members with opinions of a given individual in this group. We demonstrate that so-defined utility applied to our empirical data increases after each observed change. We then introduce an analytical model and show that it accurately recreates the trends observed in the NetSense data.

Understanding the dynamics of cooperative behavior of individuals in complex societies represents a fundamental research question which puzzles scientists working in heterogeneous fields. Many studies have been developed using the unitary agent assumption, which embeds the idea that when making decisions, individuals share the same socio-cultural parameters. In this paper, we propose the ECHO-EGN model, based on Evolutionary Game Theory, which relaxes this strong assumption by considering the heterogeneity of three fundamental socio-cultural aspects ruling the behavior of groups of people: the propensity to be more cooperative with members of the same group (Endogamic cooperation), the propensity to cooperate with the public domain (Civicness) and the propensity to prefer connections with members of the same group (Homophily). The ECHO-EGN model is shown to have high performance in describing real world behavior of interacting individuals living in complex environments. Extensive numerical experiments allowing the comparison of real data and model simulations confirmed that the introduction of the above mechanisms enhances the realism in the modelling of cooperation dynamics. Additionally, theoretical findings allow us to conclude that endogamic cooperation may limit significantly the emergence of cooperation.

Attributed network embedding aims to learn low-dimensional node vector representations in the network. To date, the primary strategy of the existing approaches is to combine topological structure with attribute information based on the homophily assumption imposed on the whole attributed networks. However, this strategy ignores the formation of community structure driven by different subspaces of attributes. In fact, (1) not all attributes are relevant to the formation of clusters; (2) the generating mechanisms of different clusters may depend on distinct relevant attributes. Therefore, we propose a new network embedding method in this paper, called Structured Subspace Embedding (SSE), to learn a compact low-dimensional vector representation. We wish to consider the cluster structure in subspaces of attributes, which reflects the different generating mechanisms of cluster structure. To this end, the attribute matrix is first represented as a latent feature representation via matrix factorization. Afterward, the local homophily constraint between network topological structure and latent feature representation is imposed. Finally, the community-selection constraint is added to further ensure the latent feature representation has a subspace cluster structure via sparse representation. Extensive experimental results on real-world networks have demonstrated its superiority on both tasks of community detection and node classification over many state-of-the-art approaches.

Adaptive voter models (AVMs) are simple mechanistic systems that model the emergence of mesoscopic structure from local networked processes driven by conflict and homophily. AVMs display rich behavior, including a phase transition from a fully-fragmented regime of "echo-chambers" to a regime of persistent disagreement governed by low-dimensional quasistable manifolds. Many extant methods for approximating the behavior of AVMs are either restricted in scope, expensive in computation, or inaccurate in predicting important statistics. In this work, we develop a novel, second-order moment closure approximation method for binary-state rewire-to-random and rewire-to-same model variants. We incorporate a small amount of noise via a random mutation term, which renders the system ergodic. Using ergodicity, we then approximate the voting process, which is non-Markovian in the second moments of the system, with a Markovian term near the phase transition. This approximation exploits an asymmetry between different classes of voting events. The resulting scheme enables us to predict the location of the phase transition and the active edge density in the regime of persistent disagreement, across the entire space of parameters and opinion densities. Numerically, our results are nearly exact for the rewire-to-random model, and competitive with other current approaches for the rewire-to-same model. Moreover, our computations display constant scaling in the mean degree, enabling approximations for denser systems than previously possible. We conclude with suggestions for model refinements and extensions., Comment: 20 Pages, 4 figures. Final revised copy

Online Social Networks (OSNs) are largely popular. People interact daily on networks such as Facebook, Twitter, YouTube, Email, Messenger, WhatsApp, Google+, Quora, LiveJournal etc. and the types of interaction on these networks are different. The types of content that people share on these networks are also different. Each of the OSNs serves a different purpose, such as sharing multimedia, microblogging, serving as a Question and Answer forum etc. Moreover, the community structure of these networks is also different. However, currently, all OSNs are considered as scale-free network based on the power-law degree distribution nature of these networks. In this paper, the effect of network properties such as density, diameter, degree distribution, global clustering coefficient, local clustering coefficient, homophily, assortativity and other centrality measures such as power-law exponent, eigenvector centrality and closeness centrality on seven online social networks, namely, Facebook, Twitter, YouTube, Email, Google+, Epinions and Gowalla networks are studied. The differences and similarities in the empirical results are analyzed, and the OSNs are divided categories based on the observed differences and similarities in network properties and centrality measures.

In this paper, a software environment to support Network-Oriented Modeling is presented. The environment has been implemented in MATLAB. This code covers the principles of temporal-causal network models. The software environment has built-in options for network adaptation principles such as the Hebbian learning principle from neuroscience and the adaptation principle for bonding based on homophily from social science. The implementation is illustrated for an adaptive temporal-causal network model under acute stress for decision-making.

Online user behaviors are increasingly modulated by social media. Extant literature mainly focuses on investigating how network structures affect user behaviors. However, recent empirical results demonstrate that user behaviors and network structures usually coevolve dynamically, and topological patterns turn out to be inadequate for characterizing real-world user behaviors. In this paper, we present a dynamic model to deal with this challenge. This proposed model is mainly governed by two competing principles: homophily and homeostasis. Empirical evaluations of three online real-world datasets suggest that the proposed dynamic model can well predict long-range online user behaviors.

The challenging but intriguing problem of modeling opinion formation dynamics in online social networks (OSNs) has attracted many researchers in recent years because the inherent complexities present in human opinion update process are yet to be clearly understood. Although the existing works adopt the distance-based homophily principle to model the neighbors’ influences on the formation of an agent’s opinion, they ignore several other key factors that govern the update process. Explicitly, we consider two essential aspects of the real-world opinion formation process that were not explored previously. First, we consider the predisposition of agents that leads to selective exposure to information when presented with different opinion sources. Second, we explicitly consider an agent’s past interaction experience with others and how opinions encountered in the past interactions influence future opinion update process of that agent. Although the confidence level of an agent on the expressed opinion was previously used to distinguish an expert, we propose the concept of the relative credibility of the opinion sources for such distinction. For this, we take into account an agent’s perceived credibility about others and the relative nature of human judgment when exposed to many opinion sources with different credibility. In addition, for the first time, the credibility of sources external to an OSN is considered in the opinion formation model proposed in this paper. We validate our model by analyzing its performance in capturing the real-world opinion formation dynamics using traces collected from an OSN, specifically Twitter. On the other hand, through simulation, various scenarios are created to observe the steady-state outcomes of the dynamics under various influences of our model parameters and network characteristics. Finally, different compelling and practical applications with social and economic values can be built based on our model.

This study develops a learning-simulation coordinated method to perform individual-level causal inference and social influence identification in social media. This method uses machine learning models to predict user adoption behavior, uses simulation to infer unobservable potential outcomes, and uses a counterfactual framework to identify individual-level social influence. The method also uses an adjusting strategy to reduce the effect of homophily and correlated unobservables. Empirical results obtained on a synthetic dataset and a semi-synthetic dataset show that the proposed method performs better on causal inference at the individual and aggregate levels than competitive methods. The computational experiment using a real-world database considers three applications, i.e., new product adoption, repeated purchase and cross selling. The empirical results show that the proposed method performs well on identifying influential members. The results reveal that the global hubs and local central nodes of the versatile friend circles have similar influences on the adoption behavior of the followers.

We consider the specification of effects of numerical actor attributes, having an interval level of measurement, in statistical models for directed social networks. A fundamental mechanism is homophily or assortativity, where actors have a higher likelihood to be tied with others having similar values of the variable under study. But there are other mechanisms that may also play a role in how the attribute values of two actors influence the likelihood of a tie between them. We discuss three additional mechanisms: aspiration, the tendency to send more ties to others having high values; attachment conformity, sending more ties to others whose values are close to the “social norm”; and sociability, where those having higher values will tend to send more ties generally. These mechanisms may operate jointly, and then their effects will be confounded. We present a specification representing these effects simultaneously by a four-parameter quadratic function of the values of sender and receiver. Flexibility can be increased by a five-parameter extension. We argue that for numerical actor attributes having important effects on directed networks, these specifications may provide an improvement. An illustration is given of dependence of advice ties on academic grades, analyzed by the Stochastic Actor-oriented Model.

Graph Neural Networks (GNNs) have achieved significant success in learning better representations by performing feature propagation and transformation iteratively to leverage neighborhood information. Nevertheless, iterative propagation restricts the information of higher-layer neighborhoods to be transported through and fused with the lower-layer neighborhoods', which unavoidably results in feature smoothing between neighborhoods in different layers and can thus compromise the performance, especially on heterophily networks. Furthermore, most deep GNNs only recognize the importance of higher-layer neighborhoods while yet to fully explore the importance of multi-hop dependency within the context of different layer neighborhoods in learning better representations. In this work, we first theoretically analyze the feature smoothing between neighborhoods in different layers and empirically demonstrate the variance of the homophily level across neighborhoods at different layers. Motivated by these analyses, we further propose a tree decomposition method to disentangle neighborhoods in different layers to alleviate feature smoothing among these layers. Moreover, we characterize the multi-hop dependency via graph diffusion within our tree decomposition formulation to construct Tree Decomposed Graph Neural Network (TDGNN), which can flexibly incorporate information from large receptive fields and aggregate this information utilizing the multi-hop dependency. Comprehensive experiments demonstrate the superior performance of TDGNN on both homophily and heterophily networks under a variety of node classification settings. Extensive parameter analysis highlights the ability of TDGNN to prevent over-smoothing and incorporate features from shallow layers with deeper multi-hop dependencies, which provides new insights towards deeper graph neural networks. Code of TDGNN: http://github.com/YuWVandy/TDGNN, CIKM 2021

Android has become the most popular mobile operating system. Correspondingly, an increasing number of Android malware has been developed and spread to steal users’ private information. There exists one type of malware whose benign behaviors are developed to camouflage malicious behaviors. The malicious component occupies a small part of the entire code of the application (app for short), and the malicious part is strongly coupled with the benign part. In this case, the malware may cause false negatives when malware detectors extract features from the entire apps to conduct classification because the malicious features of these apps may be hidden among benign features. Moreover, some previous work aims to divide the entire app into several parts to discover the malicious part. However, the premise of these methods to commence app partition is that the connections between the normal part and the malicious part are weak (repackaged malware). In this paper, we call this type of malware as Android covert malware and generate the first dataset of covert malware. To detect covert malware samples, we first conduct static analysis to extract the function call graphs. Through the deep analysis on call graphs, we observe that although the correlations between the normal part and the malicious part in these graphs are high, the degree of these correlations has a unique range of distribution. Based on the observation, we design a novel system, HomDroid, to detect covert malware by analyzing the homophily of call graphs. We identify the ideal threshold of correlation to distinguish the normal part and the malicious part based on the evaluation results on a dataset of 4,840 benign apps and 3,385 covert malicious apps. According to our evaluation results, HomDroid is capable of detecting 96.8% of covert malware while the False Negative Rates of another four state-of-the-art systems (PerDroid, Drebin, MaMaDroid, and IntDroid) are 30.7%, 16.3%, 15.2%, and 10.4%, respectively.

Teammate invitation networks are foundational for team assembly, and recommender systems (similar to dating websites, but for selecting potential teammates) can aid the formation of such networks. This paper extends Hinds, Carley, Krackhardt, and Wholey’s (2000) influential model of team member selection by incorporating online recommender systems. Exponential random graph modeling of two samples (overall N = 410; 63 teams; 1,048 invitations) shows the invitation network is predicted by online recommendations, beyond previously-established effects of prior collaboration/familiarity, skills/competence, and homophily. Importantly, online recommendations are less heeded when there is prior collaboration (effect replicates across samples). This study highlights technology-enabled team assembly from a network perspective.

Network graphs are used for high-stake decision making in medical and other contexts. For instance, graph drawings conveying relatedness can be relevant in the context of spreading diseases. Node-link diagrams can be used to visually assess the degree of homophily in a network—a condition where a presence of the link is more likely when nodes are similar. In an online experiment (N = 531), we tested how robustly laypeople can judge homophily from node-link diagrams and how variation of time constraints and layout of the diagrams affect judgments. The results showed that participants were able to give appropriate judgments. While granting more time led to better performance, the effects were small. Rather, the first seconds account for most of the information an individual can extract from the graphs. Furthermore, we showed a difference in performance between two types of layouts (bipartite and polarized). Results have consequences for communicating the degree of homophily in network graphs to the public.

Integrating different perspectives in the generation of knowledge is essential for understanding learning problems. And crowdsourcing provides new approaches to harvest distributed knowledge. However, students are more inclined to connect with people with similar characteristics, and the identity of a group has an impact on cooperative learning performance. Therefore, there is an attempt to integrate the homogeneous students in crowdsourcing and to explore the influence of student homophily knowledge generation framework through teaching experiments for more meaningful and effective participation in the process of knowledge generation. The results demonstrate that homophily can improve students' academic performance, and the knowledge generation framework in crowdsourcing has a positive effect on innovation generation.