arXiv subjects

Kristina Lerman

Publications and source records attributed to Kristina Lerman.

At least 181 records · Page 10Linked to original sources

arXiv2015

Structural Properties of Ego Networks

The structure of real-world social networks in large part determines the evolution of social phenomena, including opinion formation, diffusion of information and influence, and the spread of disease. Globally, network structure is characterized by features such as degree distribution, degree assortativity, and clustering coefficient. However, information about global structure is usually not available to each vertex. Instead, each vertex's knowledge is generally limited to the locally observable portion of the network consisting of the subgraph over its immediate neighbors. Such subgraphs, known as ego networks, have properties that can differ substantially from those of the global network. In this paper, we study the structural properties of ego networks and show how they relate to the global properties of networks from which they are derived. Through empirical comparisons and mathematical derivations, we show that structural features, similar to static attributes, suffer from paradoxes. We quantify the differences between global information about network structure and local estimates. This knowledge allows us to better identify and correct the biases arising from incomplete local information.

cs.SI↗

arXiv2014

The Impact of Network Flows on Community Formation in Models of Opinion Dynamics

We study dynamics of opinion formation in a network of coupled agents. As the network evolves to a steady state, opinions of agents within the same community converge faster than those of other agents. This framework allows us to study how network topology and network flow, which mediates the transfer of opinions between agents, both affect the formation of communities. In traditional models of opinion dynamics, agents are coupled via conservative flows, which result in one-to-one opinion transfer. However, social interactions are often non-conservative, resulting in one-to-many transfer of opinions. We study opinion formation in networks using one-to-one and one-to-many interactions and show that they lead to different community structure within the same network.

cs.SI↗

arXiv2014

Tripartite Graph Clustering for Dynamic Sentiment Analysis on Social Media

The growing popularity of social media (e.g, Twitter) allows users to easily share information with each other and influence others by expressing their own sentiments on various subjects. In this work, we propose an unsupervised \emph{tri-clustering} framework, which analyzes both user-level and tweet-level sentiments through co-clustering of a tripartite graph. A compelling feature of the proposed framework is that the quality of sentiment clustering of tweets, users, and features can be mutually improved by joint clustering. We further investigate the evolution of user-level sentiments and latent feature vectors in an online framework and devise an efficient online algorithm to sequentially update the clustering of tweets, users and features with newly arrived data. The online framework not only provides better quality of both dynamic user-level and tweet-level sentiment analysis, but also improves the computational and storage efficiency. We verified the effectiveness and efficiency of the proposed approaches on the November 2012 California ballot Twitter data.

cs.SI↗

arXiv2014

Partitioning Networks with Node Attributes by Compressing Information Flow

Real-world networks are often organized as modules or communities of similar nodes that serve as functional units. These networks are also rich in content, with nodes having distinguishing features or attributes. In order to discover a network's modular structure, it is necessary to take into account not only its links but also node attributes. We describe an information-theoretic method that identifies modules by compressing descriptions of information flow on a network. Our formulation introduces node content into the description of information flow, which we then minimize to discover groups of nodes with similar attributes that also tend to trap the flow of information. The method has several advantages: it is conceptually simple and does not require ad-hoc parameters to specify the number of modules or to control the relative contribution of links and node attributes to network structure. We apply the proposed method to partition real-world networks with known community structure. We demonstrate that adding node attributes helps recover the underlying community structure in content-rich networks more effectively than using links alone. In addition, we show that our method is faster and more accurate than alternative state-of-the-art algorithms.

cs.SI↗

arXiv2014

Rethinking Centrality: The Role of Dynamical Processes in Social Network Analysis

Many popular measures used in social network analysis, including centrality, are based on the random walk. The random walk is a model of a stochastic process where a node interacts with one other node at a time. However, the random walk may not be appropriate for modeling social phenomena, including epidemics and information diffusion, in which one node may interact with many others at the same time, for example, by broadcasting the virus or information to its neighbors. To produce meaningful results, social network analysis algorithms have to take into account the nature of interactions between the nodes. In this paper we classify dynamical processes as conservative and non-conservative and relate them to well-known measures of centrality used in network analysis: PageRank and Alpha-Centrality. We demonstrate, by ranking users in online social networks used for broadcasting information, that non-conservative Alpha-Centrality generally leads to a better agreement with an empirical ranking scheme than the conservative PageRank.

cs.SI↗

arXiv2014

Network Weirdness: Exploring the Origins of Network Paradoxes

Social networks have many counter-intuitive properties, including the "friendship paradox" that states, on average, your friends have more friends than you do. Recently, a variety of other paradoxes were demonstrated in online social networks. This paper explores the origins of these network paradoxes. Specifically, we ask whether they arise from mathematical properties of the networks or whether they have a behavioral origin. We show that sampling from heavy-tailed distributions always gives rise to a paradox in the mean, but not the median. We propose a strong form of network paradoxes, based on utilizing the median, and validate it empirically using data from two online social networks. Specifically, we show that for any user the majority of user's friends and followers have more friends, followers, etc. than the user, and that this cannot be explained by statistical properties of sampling. Next, we explore the behavioral origins of the paradoxes by using the shuffle test to remove correlations between node degrees and attributes. We find that paradoxes for the mean persist in the shuffled network, but not for the median. We demonstrate that strong paradoxes arise due to the assortativity of user attributes, including degree, and correlation between degree and attribute.

cs.SI↗

arXiv2013

LA-CTR: A Limited Attention Collaborative Topic Regression for Social Media

Probabilistic models can learn users' preferences from the history of their item adoptions on a social media site, and in turn, recommend new items to users based on learned preferences. However, current models ignore psychological factors that play an important role in shaping online social behavior. One such factor is attention, the mechanism that integrates perceptual and cognitive features to select the items the user will consciously process and may eventually adopt. Recent research has shown that people have finite attention, which constrains their online interactions, and that they divide their limited attention non-uniformly over other people. We propose a collaborative topic regression model that incorporates limited, non-uniformly divided attention. We show that the proposed model is able to learn more accurate user preferences than state-of-art models, which do not take human cognitive factors into account. Specifically we analyze voting on news items on the social news aggregator and show that our model is better able to predict held out votes than alternate models. Our study demonstrates that psycho-socially motivated models are better able to describe and predict observed behavior than models which only consider latent social structure and content.

cs.IR↗

arXiv2013

Spectral Clustering with Epidemic Diffusion

Spectral clustering is widely used to partition graphs into distinct modules or communities. Existing methods for spectral clustering use the eigenvalues and eigenvectors of the graph Laplacian, an operator that is closely associated with random walks on graphs. We propose a new spectral partitioning method that exploits the properties of epidemic diffusion. An epidemic is a dynamic process that, unlike the random walk, simultaneously transitions to all the neighbors of a given node. We show that the replicator, an operator describing epidemic diffusion, is equivalent to the symmetric normalized Laplacian of a reweighted graph with edges reweighted by the eigenvector centralities of their incident nodes. Thus, more weight is given to edges connecting more central nodes. We describe a method that partitions the nodes based on the componentwise ratio of the replicator's second eigenvector to the first, and compare its performance to traditional spectral clustering techniques on synthetic graphs with known community structure. We demonstrate that the replicator gives preference to dense, clique-like structures, enabling it to more effectively discover communities that may be obscured by dense intercommunity linking.

cs.SI↗

arXiv2013

The Simple Rules of Social Contagion

It is commonly believed that information spreads between individuals like a pathogen, with each exposure by an informed friend potentially resulting in a naive individual becoming infected. However, empirical studies of social media suggest that individual response to repeated exposure to information is significantly more complex than the prediction of the pathogen model. As a proxy for intervention experiments, we compare user responses to multiple exposures on two different social media sites, Twitter and Digg. We show that the position of the exposing messages on the user-interface strongly affects social contagion. Accounting for this visibility significantly simplifies the dynamics of social contagion. The likelihood an individual will spread information increases monotonically with exposure, while explicit feedback about how many friends have previously spread it increases the likelihood of a response. We apply our model to real-time forecasting of user behavior.

cs.SI↗

arXiv2013

Social Contagion: An Empirical Study of Information Spread on Digg and Twitter Follower Graphs

Social networks have emerged as a critical factor in information dissemination, search, marketing, expertise and influence discovery, and potentially an important tool for mobilizing people. Social media has made social networks ubiquitous, and also given researchers access to massive quantities of data for empirical analysis. These data sets offer a rich source of evidence for studying dynamics of individual and group behavior, the structure of networks and global patterns of the flow of information on them. However, in most previous studies, the structure of the underlying networks was not directly visible but had to be inferred from the flow of information from one individual to another. As a result, we do not yet understand dynamics of information spread on networks or how the structure of the network affects it. We address this gap by analyzing data from two popular social news sites. Specifically, we extract follower graphs of active Digg and Twitter users and track how interest in news stories cascades through the graph. We compare and contrast properties of information cascades on both sites and elucidate what they tell us about dynamics of information flow on networks.

cs.SI↗

arXiv2013

Stochastic Models Predict User Behavior in Social Media

User response to contributed content in online social media depends on many factors. These include how the site lays out new content, how frequently the user visits the site, how many friends the user follows, how active these friends are, as well as how interesting or useful the content is to the user. We present a stochastic modeling framework that relates a user's behavior to details of the site's user interface and user activity and describe a procedure for estimating model parameters from available data. We apply the model to study discussions of controversial topics on Twitter, specifically, to predict how followers of an advocate for a topic respond to the advocate's posts. We show that a model of user behavior that explicitly accounts for a user transitioning through a series of states before responding to an advocate's post better predicts response than models that fail to take these states into account. We demonstrate other benefits of stochastic models, such as their ability to identify users who are highly interested in advocate's posts.

cs.CY↗

arXiv2013

Attention and Visibility in an Information Rich World

As the rate of content production grows, we must make a staggering number of daily decisions about what information is worth acting on. For any flourishing online social media system, users can barely keep up with the new content shared by friends. How does the user-interface design help or hinder users' ability to find interesting content? We analyze the choices people make about which information to propagate on the social media sites Twitter and Digg. We observe regularities in behavior which can be attributed directly to cognitive limitations of humans, resulting from the different visibility policies of each site. We quantify how people divide their limited attention among competing sources of information, and we show how the user-interface design can mediate information spread.

cs.SI↗

arXiv2013

Friendship Paradox Redux: Your Friends Are More Interesting Than You

Feld's friendship paradox states that "your friends have more friends than you, on average." This paradox arises because extremely popular people, despite being rare, are overrepresented when averaging over friends. Using a sample of the Twitter firehose, we confirm that the friendship paradox holds for >98% of Twitter users. Because of the directed nature of the follower graph on Twitter, we are further able to confirm more detailed forms of the friendship paradox: everyone you follow or who follows you has more friends and followers than you. This is likely caused by a correlation we demonstrate between Twitter activity, number of friends, and number of followers. In addition, we discover two new paradoxes: the virality paradox that states "your friends receive more viral content than you, on average," and the activity paradox, which states "your friends are more active than you, on average." The latter paradox is important in regulating online communication. It may result in users having difficulty maintaining optimal incoming information rates, because following additional users causes the volume of incoming tweets to increase super-linearly. While users may compensate for increased information flow by increasing their own activity, users become information overloaded when they receive more information than they are able or willing to process. We compare the average size of cascades that are sent and received by overloaded and underloaded users. And we show that overloaded users post and receive larger cascades and they are poor detector of small cascades.

cs.SI↗

arXiv2013

Limited Attention and Centrality in Social Networks

How does one find important or influential people in an online social network? Researchers have proposed a variety of centrality measures to identify individuals that are, for example, often visited by a random walk, infected in an epidemic, or receive many messages from friends. Recent research suggests that a social media users' capacity to respond to an incoming message is constrained by their finite attention, which they divide over all incoming information, i.e., information sent by users they follow. We propose a new measure of centrality --- limited-attention version of Bonacich's Alpha-centrality --- that models the effect of limited attention on epidemic diffusion. The new measure describes a process in which nodes broadcast messages to their out-neighbors, but the neighbors' ability to receive the message depends on the number of in-neighbors they have. We evaluate the proposed measure on real-world online social networks and show that it can better reproduce an empirical influence ranking of users than other popular centrality measures.

cs.SI↗

arXiv2013

Structural and Cognitive Bottlenecks to Information Access in Social Networks

Information in networks is non-uniformly distributed, enabling individuals in certain network positions to get preferential access to information. Social scientists have developed influential theories about the role of network structure in information access. These theories were validated through numerous studies, which examined how individuals leverage their social networks for competitive advantage, such as a new job or higher compensation. It is not clear how these theories generalize to online networks, which differ from real-world social networks in important respects, including asymmetry of social links. We address this problem by analyzing how users of the social news aggregator Digg adopt stories recommended by friends, i.e., users they follow. We measure the impact different factors, such as network position and activity rate; have on access to novel information, which in Digg's case means set of distinct news stories. We show that a user can improve his information access by linking to active users, though this becomes less effective as the number of friends, or their activity, grows due to structural network constraints. These constraints arise because users in structurally diverse position within the follower graph have topically diverse interests from their friends. Moreover, though in most cases user's friends are exposed to almost all the information available in the network, after they make their recommendations, the user sees only a small fraction of the available information. Our study suggests that cognitive and structural bottlenecks limit access to novel information in online social networks.

cs.SI↗

arXiv2013

LA-LDA: A Limited Attention Topic Model for Social Recommendation

Social media users have finite attention which limits the number of incoming messages from friends they can process. Moreover, they pay more attention to opinions and recommendations of some friends more than others. In this paper, we propose LA-LDA, a latent topic model which incorporates limited, non-uniformly divided attention in the diffusion process by which opinions and information spread on the social network. We show that our proposed model is able to learn more accurate user models from users' social network and item adoption behavior than models which do not take limited attention into account. We analyze voting on news items on the social news aggregator Digg and show that our proposed model is better able to predict held out votes than alternative models. Our study demonstrates that psycho-socially motivated models have better ability to describe and predict observed behavior than models which only consider topics.

cs.SI↗

arXiv2012

How Visibility and Divided Attention Constrain Social Contagion

How far and how fast does information spread in social media? Researchers have recently examined a number of factors that affect information diffusion in online social networks, including: the novelty of information, users' activity levels, who they pay attention to, and how they respond to friends' recommendations. Using URLs as markers of information, we carry out a detailed study of retweeting, the primary mechanism by which information spreads on the Twitter follower graph. Our empirical study examines how users respond to an incoming stimulus, i.e., a tweet (message) from a friend, and reveals that %retweeting behavior is constrained by a few simple principles. the "principle of least effort" combined with limited attention plays a dominant role in retweeting behavior. Specifically, we observe that users retweet information when it is most visible, such as when it near the top of their Twitter stream. Moreover, our measurements quantify how a user's limited attention is divided among incoming tweets, providing novel evidence that highly connected individuals are less likely to propagate an arbitrary tweet. Our study indicates that the finite ability to process incoming information constrains social contagion, and we conclude that rapid decay of visibility is the primary barrier to information propagation online.

physics.soc-ph↗

arXiv2012

Impact of Dynamic Interactions on Multi-Scale Analysis of Community Structure in Networks

To find interesting structure in networks, community detection algorithms have to take into account not only the network topology, but also dynamics of interactions between nodes. We investigate this claim using the paradigm of synchronization in a network of coupled oscillators. As the network evolves to a global steady state, nodes belonging to the same community synchronize faster than nodes belonging to different communities. Traditionally, nodes in network synchronization models are coupled via one-to-one, or conservative interactions. However, social interactions are often one-to-many, as for example, in social media, where users broadcast messages to all their followers. We formulate a novel model of synchronization in a network of coupled oscillators in which the oscillators are coupled via one-to-many, or non-conservative interactions. We study the dynamics of different interaction models and contrast their spectral properties. To find multi-scale community structure in a network of interacting nodes, we define a similarity function that measures the degree to which nodes are synchronized and use it to hierarchically cluster nodes. We study real-world social networks, including networks of two social media providers. To evaluate the quality of the discovered communities in a social media network we propose a community quality metric based on user activity. We find that conservative and non-conservative interaction models lead to dramatically different views of community structure even within the same network. Our work offers a novel mathematical framework for exploring the relationship between network structure, topology and dynamics.

cs.SI↗