Abstract: We introduce a stochastic graph-based method for computing relative importance of textual units for Natural Language Processing. We test the technique on the problem of Text Summarization (TS). Extractive TS relies on the concept of sentence salience to identify the most important sentences in a document or set of documents. Salience is typically defined in terms of the presence of particular important words or in terms of similarity to a centroid pseudo-sentence. We consider a new approach, LexRank, for computing sentence importance based on the concept of eigenvector centrality in a graph representation of sentences. In this model, a connectivity matrix based on intra-sentence cosine similarity is used as the adjacency matrix of the graph representation of sentences. Our system, based on LexRank ranked in first place in more than one task in the recent DUC 2004 evaluation. In this paper we present a detailed analysis of our approach and apply it to a larger data set including data from earlier DUC evaluations. We discuss several methods to compute centrality using the similarity graph. The results show that degree-based methods (including LexRank) outperform both centroid-based methods and other systems participating in DUC in most of the cases. Furthermore, the LexRank with threshold method outperforms the other degree-based techniques including continuous LexRank. We also show that our approach is quite insensitive to the noise in the data that may result from an imperfect topical clustering of documents.

Abstract: We investigate random walks on complex networks and derive an exact expression for the mean firstpassage time (MFPT) between two nodes. We introduce for each node the random walk centrality C, which is the ratio between its coordination number and a characteristic relaxation time, and show that it determines essentially the MFPT. The centrality of a node determines the relative speed by which a node can receive and spread information over the network in a random process. Numerical simulations of an ensemble of random walkers moving on paradigmatic network models confirm this analytical prediction.

Abstract: We investigate random walks on complex networks and derive an exact expression for the mean firstpassage time (MFPT) between two nodes. We introduce for each node the random walk centrality C, which is the ratio between its coordination number and a characteristic relaxation time, and show that it determines essentially the MFPT. The centrality of a node determines the relative speed by which a node can receive and spread information over the network in a random process. Numerical simulations of an ensemble of random walkers moving on paradigmatic network models confirm this analytical prediction.

Abstract: Drawing on social exchange and similarity-attraction theories, we hypothesized that individuals' demographic characteristics, values, and personality influence their acquisition of central positions in their teams' social networks. Education and neuroticism predicted centrality five months later; individuals who were highly educated and low in neuroticism became high in advice and friendship centrality and low in adversarial centrality. Team members' values similarity to their teammates also predicted advice and friendship centrality; demographic similarity had limited effects.

Abstract: The current study extends research in the area of identity conflict or interference by focusing on a new identity combination, the woman and scientist identities. In addition, it examines the influence of identity centrality, or importance, as a predictor of interference and moderator of the relation between interference and well-being and science performance. Supporting hypotheses, greater identity interference was related to lower levels of performance and well-being. Furthermore, woman centrality was unrelated to interference for those with a central scientist identity, but for those without a central scientist identity, they were positively related. Although central identities were related to positive outcomes in the absence of interference, the outcomes of all women suffered when interference was high, contrary to the hypothesis. The implications of identity centrality for understanding the negotiation of potentially conflicting identities, and for the retention of women in the sciences, are discussed.

Abstract: This paper proposes a novel generalization model for selecting characteristic streets in an urban street network. This model retains the central structure of a street network. It relies on a structural representation of a street network using graph principles where vertices represent named streets and links represent street intersections. Based on this representation, so-called connectivity graph, centrality measures are introduced to qualify the status of each individual vertex within the graph. We show that these measures can be used for characterizing the structural properties of an urban street network, and for the selection of important streets. The proposed approach is validated by a case study applied to a middle-sized Swedish city.

Abstract: Community structures are an important feature of many social, biological, and technological networks. Here we study a variation on the method for detecting such communities proposed by Girvan and Newman and based on the idea of using centrality measures to define the community boundaries [M. Girvan and M. E. J. Newman, Proc. Natl. Acad. Sci. U.S.A. 99, 7821 (2002)]. We develop an algorithm of hierarchical clustering that consists in finding and removing iteratively the edge with the highest information centrality. We test the algorithm on computer generated and real-world networks whose community structure is already known or has been studied by means of other methods. We show that our algorithm, although it runs to completion in a time O(n4) , is very effective especially when the communities are very mixed and hardly detectable by the other methods.

Abstract: Many definitions of forgiveness currently exist in the literature. The current research adds to this discussion by utilizing a prototype approach to examine lay conceptions of forgiveness. A prototype approach involves categorizing objects or events in terms of their similarity to a good example, whereas a classical approach requires that there are essential elements that must be present. In Study 1, participants listed the features of forgiveness. Study 2 obtained centrality ratings for these features. In Studies 3 and 4, central features were found to be more salient in memory than peripheral features. Study 5 showed that feature centrality influenced participants’ ratings of victims involved in hypothetical transgressions. Thus, the two criteria for demonstrating prototype structure (that participants find it meaningful to judge features in terms of their centrality and that centrality affects cognition) were met.

Abstract: Social studies researchers use graphs to model group activities in social networks. An important property in this context is the centrality of a vertex: the inverse of the average distance to each other vertex. We describe a randomized approximation algorithm for centrality in weighted graphs. For graphs exhibiting the small world phenomenon, our method estimates the centrality of all vertices with high probability within a (1+epsilon) factor in near-linear time.

Abstract: The importance of assessment and the centrality of students within the process are emphasised. The changing role of the student is discussed and a number of suggestions made on how to involve learn...

Abstract: Recreation specialization theory predicts that individuals will differ in their physical, management, and social setting preferences. Few studies, however, support the hypothesis that individuals choose recreation settings consistent with their level of specialization. This study examined the association between behavioral, cognitive, and affective dimensions of specialization and site choice among vehicle-based campers in Alberta, Canada. Data were collected using on-site interviews and a mail survey. Campers at unmanaged sites (no facilities and services) had higher centrality scores, had greater familiarity with the site and more experience with unmanaged sites, and a higher level of bush skill than campers at managed sites. An ordered multinomial logit model showed that the more familiar individuals were with the site and campground type, the higher the level of bush skill, and the more important and central camping was in an individual's life, the greater the probability of choosing a campground type...

Abstract: We investigate the properties of the spanning trees of various real-world and model networks. The spanning tree representing the communication kernel of the original network is determined by maximizing the total weight of the edges, whose weights are given by the edge betweenness centralities. We find that a scale-free tree and shortcuts organize a complex network. Especially, in ubiquitous scale-free networks, it is found that the scale-free spanning tree shows very robust betweenness centrality distributions and the remaining shortcuts characterize the properties of the original network, such as the clustering coefficient and the classification of scale-free networks by the betweenness centrality distribution.

Abstract: Individuals who had won the lottery responded to a survey concerning whether they had continued to work after winning. They were also asked to indicate how important work was in their life using items and scales commonly used to measure work centrality. The authors predicted that whether lottery winners would continue to work would be related to their level of work centrality as well as to the amount of their winnings. Individuals who won large amounts in the lottery would be less likely to quit work if they had relatively greater degrees of work centrality. After controlling for a number of variables (i.e., age, gender, education, occupation, and job satisfaction), results indicated that work centrality and the amount won were significantly related to whether individuals continued to work and, as predicted, the interaction between the two was also significantly related to work continuance.

Abstract: Dynamic criminal network analysis is important for national security but also very challenging. However, little research has been done in this area. In this paper we propose to use several descriptive measures from social network analysis research to help detect and describe changes in criminal organizations. These measures include centrality for individuals, and density, cohesion, and stability for groups. We also employ visualization and animation methods to present the evolution process of criminal networks. We conducted a field study with several domain experts to validate our findings from the analysis of the dynamics of a narcotics network. The feedback from our domain experts showed that our approaches and the prototype system could be very helpful for capturing the dynamics of criminal organizations and assisting crime investigation and criminal prosecution.

Abstract: In this study, we examine how the characteristics of clique structures affect the performance of firms embedded within the cliques. Although it is generally accepted in organization theory and strategic management that firms are embedded within ego and overarching industry networks that each affect their behavior and performance, there is little evidence on whether cliques are stable features of industry networks or affect firm behavior or performance. We theorize that the value of a clique to its members depends on (1) the network centrality of the clique and (2) the internal structure and organization (heterogeneity and inequality) of the clique. Our analysis of the Canadian investment banking industry from 1952 to 1990 provides empirical evidence of stable cliques, and indicates that while a clique's internal structure and organization materially affect firm-level benefits of clique membership, its positional embeddedness within the industry network does not. Copyright © 2004 John Wiley & Sons, Ltd.

Abstract: We introduce a new measure of centrality, the information centrality C^I, based on the concept of efficient propagation of information over the network. C^I is defined for both valued and non-valued graphs, and applies to groups and classes as well as individuals. The new measure is illustrated and compared to the standard centrality measures by using a classic network data set.

Abstract: The desires to live in another country or to emigrate from one's country of origin was examined in a sample of 3200 university students from Croatia, the Czech Republic, Poland, Russia, and Slovenia. All of these countries have been experiencing economic difficulties during their transition from socialist to market-driven economies. It was hypothesized that students who wanted to emigrate would score higher in Achievement and Power Motivation and would also show higher levels of Work Centrality and lower levels of Family Centrality than those who wanted to stay in their country of origin. Motive predictors were further expected to be most important for those with high Work Centrality. As predicted, high Work Centrality and low Family Centrality were found to differ for those who wanted to leave as compared to those who wished to remain in their country. The predicted interactions for motivation and Work Centrality were supported. Achievement Motive levels alone did not relate to emigration desires, but Power Motivation did differ for the two groups, as predicted.

Abstract: Social network analysis is a set of procedures-mathematical and graphical techniques-that use indices of relatedness among entities to represent social structures in a compact and systematic manner There are several general goals of network analysis The first goal is to represent relationships of interest visually as a network or graph, and to display information in a way that allows the user to see relationships among actors embedded in the overall network A second goal is to examine basic properties of relationships in a network, such as density, centrality, and prestige A third goal is to test hypotheses regarding the structure of connections among actors Social network analysts can examine the effects that relationships have on constraining or enhancing individual behavior or network efficiency A major advantage of the network approach is that it focuses on the relationships among actors embedded in their social context

Abstract: This paper presents research on the use of household lists. Drawing on an ethnographic study of mothers' work, it focuses on the centrality of paper lists in home- and child-care arrangements, and reveals that they provide a useful means for organizing the complex interrelations between a household's people, activities and tasks. However, paper lists are also shown to be poor at handling the separation, or classification, of these things. In conclusion, both these positive and negative aspects of list making are used to raise broad pointers for CSCW and system design.

Abstract: Criminals are embedded in a network of relationships. Social ties among criminals are modeled by means of a graph where criminals compete for a booty and benefit from local interactions with their neighbours. Each criminal decides in a non-cooperative way how much crime effort he will exert. We show that the Nash equilibrium crime effort of each individual is proportional to his equilibrium Bonacich-centrality in the network, thus establishing a bridge to the sociology literature on social networks. We then analyze a policy that consists of finding and getting rid of the key player, that is, the criminal who, once removed, leads to the maximum reduction in aggregate crime. We provide a geometric characterization of the key player identified with an optimal inter-centrality measure, which takes into account both a player’s centrality and his contribution to the centrality of the others. We also provide a geometric characterization of the key group, which generalizes the key player for a group of criminals of a given size. We finally endogeneize the crime participation decision, resulting in a key player policy, which effectiveness depends on the outside opportunities available to criminals.

Abstract: Criminals are embedded in a network of relationships. Social ties among criminals are modeled by means of a graph where criminals compete for a booty and benefit from local interactions with their neighbours. Each criminal decides in a non-cooperative way how much crime effort he will exert. We show that the Nash equilibrium crime effort of each individual is proportional to his equilibrium Bonacich-centrality in the network, thus establishing a bridge to the sociology literature on social networks. We then analyze a policy that consists of finding and getting rid of the key player, that is, the criminal who, once removed, leads to the maximum reduction in aggregate crime. We provide a geometric characterization of the key player identified with an optimal inter-centrality measure, which takes into account both a player's centrality and his contribution to the centrality of the others. We also provide a geometric characterization of the key group, which generalizes the key player for a group of criminals of a given size. We finally endogeneize the crime participation decision, resulting in a key player policy, which effectiveness depends on the outside opportunities available to criminals.