### Citations

3619 | Authoritative sources in a hyperlinked environment
- Kleinberg
- 1999
(Show Context)
Citation Context ... dominate the results of community detection. 72 An intuitive method to distinguish kernel members from others is to first perform a link analysis algorithm (e.g., degree ranking, PageRank [38], HITS =-=[39]-=-) on the network to find the “influential” vertices, and then apply a cutor conductance-based community detection algorithm to those vertices only. In this way, we obtain communities solely based on t... |

3259 | The PageRank citation ranking: Bringing order to the web
- Page, Brin, et al.
- 1998
(Show Context)
Citation Context ...el ones may dominate the results of community detection. 72 An intuitive method to distinguish kernel members from others is to first perform a link analysis algorithm (e.g., degree ranking, PageRank =-=[38]-=-, HITS [39]) on the network to find the “influential” vertices, and then apply a cutor conductance-based community detection algorithm to those vertices only. In this way, we obtain communities solely... |

1992 |
Reducibility among combinatorial problems, in
- Karp
- 1972
(Show Context)
Citation Context ...L( w) = ∑ (u,v)∈E w(u) ·sw(v) subject to ∑ v∈V wi(v) = k, ∀i ∈ {1, · · · ,s}; ∑ 1i wi(v) 1, ∀v ∈ V ; wi(v) 0, ∀v ∈ V, ∀i ∈ {1, · · · ,s}. (3.2) Solving this optimization problem is intractable =-=[40]-=-. Thus, we approximate the solution by iteratively solving its one-dimensional version L(w). For each detected kernel, we give the following theorem: Theorem 3.3.1. A global maximum of the objective f... |

1510 |
Community structure in social and biological networks. The National Academy of Sciences of the United
- Girvan, Newman
(Show Context)
Citation Context ...rience that communities exist in these networks [27]. Community was often considered to be a subset of vertices that are densely connected internally but sparsely connected to the rest of the network =-=[4, 11, 15, 23, 27]-=-. For example, Newman constructed the measure of betweenness and modularity to partition a social network into disjoint communities [4,11]. Andersen et al. [15] proposed a local graph partitioning alg... |

1475 | Finding and evaluating community structure in networks - Newman, Girvan |

1185 | A fast and high quality multilevel scheme for partitioning irregular graphs - Karypis, Kumar - 1998 |

986 | Maximizing the spread of influence through a social network
- Kempe, Kleinberg, et al.
- 2003
(Show Context)
Citation Context ...ase, are a pervasive phenomenon in many networks. Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks =-=[43, 44]-=-, viral marketing [45, 46], epidemiology [47], and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal... |

975 | Modularity and community structure in networks - Newman - 2006 |

687 |
Finding community structure in very large networks
- Clauset, Newman, et al.
(Show Context)
Citation Context ...kernels, i.e. K = {K1, · · · ,K}? Our problem formulation is very different from previous work on community detection. Many algorithms have been proposed for detecting communities in social networks =-=[4, 11, 14, 26, 27]-=-, however they ignore the difference among vertices and links. Thus, these algorithms fail to distinguish community ker70 nels from their auxiliary communities. In addition, Ahn et al. [36] categorize... |

567 |
The complexity of satisfiability problems
- Schaefer
- 1978
(Show Context)
Citation Context ...e has at least one true literal and at least one false literal (i.e. literals in each clause are not all equal). Then, we have the following well-known theorem: Theorem 2.1.9. NAE-3-SAT isNP-complete =-=[25]-=-. Now, define WHISKER as the problem of determining whether there exists a whisker in a given weighted undirected graph. We will formally prove that WHISKER is an NP-complete problem by constructing a... |

564 |
Fast algorithm for detecting community structure in networks
- Newman
(Show Context)
Citation Context ... network. The left figure shows the original Twitter network (three entertainers and two politicians with their followers), the middle figure shows the five communities detected by Newman’s algorithm =-=[4]-=-, and the right figure shows two community kernels and their corresponding auxiliary communities detected by our algorithm WEBA. . . . . . . . . . . . . . . . . . . . . . . . . . 66 3.2 Efficiency com... |

542 |
Statistical Models and Methods for Lifetime Data
- Lawless, Lawless
- 1982
(Show Context)
Citation Context ...cade Model We first recall some standard notation from previous literature, and then define our feature-enhanced models based on generalized cascades. Recap. Recall the standard notation from [2] and =-=[49]-=-. Given that node j was infected at time tj , the survival function of edge (j, k) is the probability that, by time tk, node k was not infected by node j. That is, S (tk|tj;αjk) = 1− F (tk|tj;αjk) , (... |

534 | Graphs over time: densification laws, shrinking diameters and possible explanations
- Leskovec, Kleinberg, et al.
- 2005
(Show Context)
Citation Context ...52 63 73 81 88 101 146 182 223 59 Citation Graph The Citation dataset was crawled from the e-print arXiv that contains 421,578 citation links among a collection of 34,546 papers in the hep-ph archive =-=[33, 34]-=-. If paper i cites paper j or vice versa, then there is an undirected edge between vertex i and vertex j in the corresponding graph. This dataset was originally released in the KDD Cup 2003 [33], and ... |

494 | Finding Community Structure in Networks Using the Eigenvectors of Matrices
- Newman
(Show Context)
Citation Context ...kernels, i.e. K = {K1, · · · ,K}? Our problem formulation is very different from previous work on community detection. Many algorithms have been proposed for detecting communities in social networks =-=[4, 11, 14, 26, 27]-=-, however they ignore the difference among vertices and links. Thus, these algorithms fail to distinguish community ker70 nels from their auxiliary communities. In addition, Ahn et al. [36] categorize... |

444 | N-gram-based text categorization
- Cavnar, Trenkle
- 1994
(Show Context)
Citation Context ...et to its language. We define a distance function with respect to language dL(fi, fj) = ⎧⎪⎨ ⎪⎩ 0,s(fi) =s(fj); 1,s(fi) =s(fj). The language information is computed using the n-gram model proposed in =-=[52]-=-. Note that this language identification algorithm provides noisy estimates. Pairwise similarity. We include pairwise similarity (a.k.a. Jaccard index) as another distance metric in our models. Give... |

367 | Mining knowledge-sharing sites for viral marketing
- Richardson, Domingos
- 2002
(Show Context)
Citation Context ...menon in many networks. Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks [43, 44], viral marketing =-=[45, 46]-=-, epidemiology [47], and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal is to devise intervention... |

304 | Graph structure in the web
- Broder, Kumar, et al.
(Show Context)
Citation Context ...re a total of 465,023 vertices in the Twitter graph, and 38,913 vertices (8%) in its strongly connected component (SCC). It is well-known that many real-world social networks have a bow-tie structure =-=[41]-=-. There is a directed path from each vertex of the set IN to (all the vertices of) SCC. Similarly, there is a directed path from (all the vertices of) SCC to each vertex of the set OUT. Fig. 3.6 shows... |

266 | Graph evolution: Densification and shrinking diameters
- Leskovec, Kleinberg, et al.
- 2007
(Show Context)
Citation Context ...1 53 62 72 85 97 148 197 244 Coauthor Graph The Coauthor dataset was crawled from the e-print arXiv that contains scientific coauthorship between authors of the papers submitted to the hep-ph archive =-=[32]-=-. If author i coauthors a paper with author j, there is an undirected edge between vertex i and vertex j in the corresponding graph. If a paper has k authors, then there is a clique of size k in the g... |

243 | Statistical properties of community structure in large social and information networks
- Leskovec, Lang, et al.
- 2008
(Show Context)
Citation Context ... corresponding graph into disjoint communities [4, 11, 15, 23, 26–30]. Conductance was often taken as the measure of the quality of community, and algorithms were sometimes restricted to dense graphs =-=[13, 22, 24, 27]-=-. However, to identify well-defined communities in social networks, one needs to realize that an individual may belong to multiple communities at the same time, and is likely to havemore connections t... |

213 |
Graph clustering
- Schaeffer
- 2007
(Show Context)
Citation Context ... S and Sc. Out of numerous densitybased measures, conductance has been extensively employed for community detection, which intends to maximize internal connectivity and minimize external connectivity =-=[22, 24]-=-. The concept of whiskers was informally introduced in [21] referring to weakly-connected subsets linked to the rest of the graph by just a single edge. Empirically, whiskers are peripheral and can be... |

207 | Community structure in large networks: Natural cluster sizes and the absence of large well-defined clusters. Internet Mathematics
- Leskovec, Lang, et al.
- 2009
(Show Context)
Citation Context ...of identifying and evaluating close-knit communities in large complex networks, most of which is based on the premise that it is a matter of common experience that communities exist in these networks =-=[21]-=-. In particular, as the Internet has become an indispensable part of our life, understanding community structure is not only crucial for studying real-world societies, but also helpful to improve the ... |

200 |
SciPy: Open source scientific tools for Python,
- Jones, Oliphant, et al.
- 2001
(Show Context)
Citation Context ...ency (i.e., elapsed time required for obtaining the optimum) of our feature-enhanced models. All algorithms are implemented using Python with the Fortran implementation of L-BFGS-B available in Scipy =-=[53]-=-, and all experiments are performed on a machine running CentOS Linux with a 6-core Intel x5690 3.46GHZ CPU and 48GB memory. 4.4.2 Quantitative Performance We trace the propagation of a set of 500 Has... |

168 | Empirical comparison of algorithms for network community detection
- Leskovec, Lang, et al.
- 2010
(Show Context)
Citation Context ...ided a novel perspective for finding hierarchical community structure by categorizing links instead of vertices. A range of community detection methods have been empirically evaluated and compared in =-=[59]-=-. Further, community detection problem has been extended to handle query-dependent cases [60]. Many studies combined link and content information for finding meaningful communities [61, 62]. The dynam... |

165 | A spectral clustering approaches to finding communities in graphs
- White, Smyth
- 2005
(Show Context)
Citation Context ... Detailed Introduction. The problem of community detection has been extensively studied and many algorithms have been proposed, such as cut- and conductance-based methods [11–14], spectral clustering =-=[4, 15, 16]-=-, (α, β)-clustering [10, 17], and topic modeling methods [18]. The cut- and conductance-based and spectral clustering methods are usually based on a fundamental assumption that communities have dense ... |

163 | Feedback effects between similarity and social influence in online communities
- Crandall, Cosley, et al.
- 2008
(Show Context)
Citation Context ... to better evaluate the quality of a community [71, 72]. Various techniques have been proposed for identifying and modeling social influence in large real-world networks. For example, Crandall et al. =-=[74]-=- studied the interactions between social influence and selection, Tang et al. [75] analyzed topic-level social influence in large-scale networks, and Gomez-Rodriguez et al. [19] developed a method to ... |

151 | Social influence analysis in large-scale networks
- Tang, Sun, et al.
- 2009
(Show Context)
Citation Context ...een proposed for identifying and modeling social influence in large real-world networks. For example, Crandall et al. [74] studied the interactions between social influence and selection, Tang et al. =-=[75]-=- analyzed topic-level social influence in large-scale networks, and Gomez-Rodriguez et al. [19] developed a method to trace paths of influence and diffusion through networks. However, most existing wo... |

147 |
Algorithm 778. L-BFGS-B, Fortran subroutines for Large-Scale bound constrained optimization
- Zhu, Byrd, et al.
- 1997
(Show Context)
Citation Context ...m. However, regular packages such as CVXOPT [50] could not handle the scale of our Twitter dataset and ran out of memory. Thus, we use the limitedmemory BFGS algorithm with box constraints (L-BFGS-B) =-=[51]-=- to solve Eq. (4.9) and Eq. (4.10) by implicitly approximating the inverse Hessian matrix. We use the box constraints to enforce the non-negativity of the transmission rates. 111 4.4 Experimental Resu... |

143 | Influentials, networks, and public opinion formation.
- Watts, Dodds
- 2007
(Show Context)
Citation Context ...menon in many networks. Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks [43, 44], viral marketing =-=[45, 46]-=-, epidemiology [47], and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal is to devise intervention... |

133 | Who says what to whom on Twitter
- Wu, Hofman, et al.
- 2011
(Show Context)
Citation Context ...hat exhibit different influence and different behavior. For instance, statistics have shown that less than 1% of the Twitter users (e.g. entertainers, politicians, writers) produce 50% of its content =-=[1]-=-, while the others (e.g. fans, followers, readers) have much less influence and completely different social behavior. In this thesis, we define and explore a novel problem called community kernel dete... |

118 |
Tracking information epidemics in blogspace
- Adar, Adamic
- 2005
(Show Context)
Citation Context ...puter virus or an infectious disease, are a pervasive phenomenon in many networks. Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion =-=[2, 19, 20, 42]-=-, social networks [43, 44], viral marketing [45, 46], epidemiology [47], and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading... |

116 | Inferring networks of diffusion and influence.
- Gomez-Rodriguez, Leskovec, et al.
- 2010
(Show Context)
Citation Context ...e network structure? This type of latent network inference problem based on the time-stamps of infection (or, information-reproduction) events has received increasing interest over the past few years =-=[2, 19, 20]-=-. Previous work was largely based on two ma13 jor assumptions: (1) the diffusion process is causal (i.e., not affected by events in the future), and (2) infection events closer in time are more likely... |

96 |
An information-theoretic framework for resolving community structure in complex networks
- Rosvall, Bergstrom
- 2007
(Show Context)
Citation Context ...kernels, i.e. K = {K1, · · · ,K}? Our problem formulation is very different from previous work on community detection. Many algorithms have been proposed for detecting communities in social networks =-=[4, 11, 14, 26, 27]-=-, however they ignore the difference among vertices and links. Thus, these algorithms fail to distinguish community ker70 nels from their auxiliary communities. In addition, Ahn et al. [36] categorize... |

95 |
Different epidemic curves for severe acute respiratory syndrome reveal similar impacts of control measures
- Wallinga, Teunis
(Show Context)
Citation Context ... Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks [43, 44], viral marketing [45, 46], epidemiology =-=[47]-=-, and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal is to devise intervention strategies to eith... |

70 |
communities reveal multi-scale complexity in networks
- Link
- 2009
(Show Context)
Citation Context ...11, 14, 26, 27], however they ignore the difference among vertices and links. Thus, these algorithms fail to distinguish community ker70 nels from their auxiliary communities. In addition, Ahn et al. =-=[36]-=- categorized links instead of vertices to discover hierarchical community structure. Mishra et al. [10] proposed the concept of (α, β)-community to allow communities to overlap. However, these algorit... |

67 | Randomization tests for distinguishing social influence and homophily effects - Fond, Neville |

62 | Community evolution in dynamic multi-mode networks - Tang, Liu, et al. - 2008 |

57 | Predicting information seeker satisfaction in community question answering. - Liu, Bian, et al. - 2008 |

56 | Uncovering the Temporal Dynamics of Diffusion Networks.
- Gomez-Rodriguez, Balduzzi, et al.
- 2011
(Show Context)
Citation Context ...aximum values for each metric are marked bold. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 3.5 Efficiency comparison on Twitter, Coauthor, and Wikipedia. . . . 95 4.1 Parametric Models =-=[2]-=-. . . . . . . . . . . . . . . . . . . . . . . . . . 104 4.2 Performance comparison on Twitter (non-splitting/exponential). 118 4.3 Performance comparison on Twitter (splitting/exponential). . . . 118 ... |

55 | On the convexity of latent social network inference. In:
- Myers, Leskovec
- 2010
(Show Context)
Citation Context ...nd remove all other duplicates from each cascade. We use the true number of edges as an input parameter for NETINF. Due to license issues with the optimization software, we do not compare with CONNIE =-=[20]-=- in this chapter, but its performance is comparable with that of NETRATE and NETINF according to previous literature. Quantitative performance. We use precision, recall, and F1-score to evaluate the... |

55 | k-core decomposition: a tool for the visualization of large scale networks. - JI, Dall’Asta, et al. - 2005 |

54 | Combining link and content for community detection: a discriminative approach
- Yang, Jin, et al.
- 2009
(Show Context)
Citation Context ...nd compared in [59]. Further, community detection problem has been extended to handle query-dependent cases [60]. Many studies combined link and content information for finding meaningful communities =-=[61, 62]-=-. The dynamic behavior of communities was also extensively explored in previous 123 work [63–66]. Other models have been proposed to improve the accuracy of community detection in different scenarios ... |

50 | Scalable graph clustering using stochastic flows: applications to community discovery - Satuluri, Parthasarathy - 2009 |

44 |
Finding effectors in social networks
- Lappas, Terzi, et al.
- 2010
(Show Context)
Citation Context ...ase, are a pervasive phenomenon in many networks. Diffusion and propagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks =-=[43, 44]-=-, viral marketing [45, 46], epidemiology [47], and ecology [48]. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal... |

43 | Inferring relevant social networks from interpersonal communication. - Choudhury, Mason, et al. - 2010 |

36 | Link prediction via matrix factorization
- McAuley, Menon, et al.
- 2011
(Show Context)
Citation Context ...ld cascades also present a range of new opportunities to define richer probabilistic models. Previous work combined latent features with explicit ones 126 to solve structural link prediction problems =-=[76]-=-. In Chapter 4, we propose a feature-enhanced framework to address the scenario where nodes can be repeatedly infected. We develop a family of novel probabilistic models based not only on the time int... |

35 | Metafac: Community discovery via relational hypergraph factorization. - Lin, Sun, et al. - 2009 |

34 | Generalized cores - Batagelj, Zaveršnik |

34 | On community outliers and their efficient detection in information networks.
- Gao, Liang, et al.
- 2010
(Show Context)
Citation Context ...nd compared in [59]. Further, community detection problem has been extended to handle query-dependent cases [60]. Many studies combined link and content information for finding meaningful communities =-=[61, 62]-=-. The dynamic behavior of communities was also extensively explored in previous 123 work [63–66]. Other models have been proposed to improve the accuracy of community detection in different scenarios ... |

34 | Sampling community structure
- Maiya, Berger-Wolf
(Show Context)
Citation Context ...2]. For example, Zhang et al. [67] proposed a novel community detection algorithm that employs a dynamic process by contradicting network topology and topologybased propinquity. Maiya and Berger-Wolf =-=[69]-=- utilized a novel method based on expander graphs to sample communities in networks. Yang et al. [73] explored a dynamic stochastic block model for finding communities and their evolution in dynamic s... |

33 | Pet: a statistical model for popular events tracking in social communities
- Lin, Zhao, et al.
- 2010
(Show Context)
Citation Context ...extensively studied and many algorithms have been proposed, such as cut- and conductance-based methods [11–14], spectral clustering [4, 15, 16], (α, β)-clustering [10, 17], and topic modeling methods =-=[18]-=-. The cut- and conductance-based and spectral clustering methods are usually based on a fundamental assumption that communities have dense internal connections and sparse external connections. (α, β)-... |

33 | Detecting communities in social networks using max-min modularity
- Chen, Zaïane, et al.
- 2009
(Show Context)
Citation Context ...63–66]. Other models have been proposed to improve the accuracy of community detection in different scenarios [67–70]. New measures have also been proposed to better evaluate the quality of community =-=[71, 72]-=-. For example, Zhang et al. [67] proposed a novel community detection algorithm that employs a dynamic process by contradicting network topology and topologybased propinquity. Maiya and Berger-Wolf [6... |

33 | A syntactic tree matching approach to finding similar questions in community-based qa services - Wang, Ming, et al. - 2009 |

27 |
Fast unfolding of community hierarchies in large networks
- Blondel, Guillaume, et al.
(Show Context)
Citation Context ...ts. This algorithm first partitions the graph into two equal-sized subgraphs, and then finds the cut with the lowest conductance whose smaller side is contained in one of the two subgraphs. LOUVAIN =-=[37]-=-: community detection algorithm based on modularity. This algorithm is in general a greedy optimization method. NEWMAN1 [11]: community detection algorithm based on betweenness. This algorithm is in... |

26 |
S.: A flow-based method for improving the expansion or conductance of graph cuts
- Lang, Rao
- 2004
(Show Context)
Citation Context ... corresponding graph into disjoint communities [4, 11, 15, 23, 26–30]. Conductance was often taken as the measure of the quality of community, and algorithms were sometimes restricted to dense graphs =-=[13, 22, 24, 27]-=-. However, to identify well-defined communities in social networks, one needs to realize that an individual may belong to multiple communities at the same time, and is likely to havemore connections t... |

25 | Constant-factor approximation algorithms for identifying dynamic communities - Tantipathananandh, Berger-Wolf - 2009 |

24 | Maximizing the spread of cascades using network design.
- Sheldon, Dilkina, et al.
- 2012
(Show Context)
Citation Context ...pagation processes have been studied in a broad range of disciplines, such as information diffusion [2, 19, 20, 42], social networks [43, 44], viral marketing [45, 46], epidemiology [47], and ecology =-=[48]-=-. In previous work, researchers have mostly focused on a number of optimization problems derived from cascading processes, where the goal is to devise intervention strategies to either maximize (e.g.,... |

24 | The community-search problem and how to plan a successful cocktail party
- Sozio, Gionis
- 2010
(Show Context)
Citation Context ...nstead of vertices. A range of community detection methods have been empirically evaluated and compared in [59]. Further, community detection problem has been extended to handle query-dependent cases =-=[60]-=-. Many studies combined link and content information for finding meaningful communities [61, 62]. The dynamic behavior of communities was also extensively explored in previous 123 work [63–66]. Other ... |

21 | Identifying temporal patterns and key players in document collections,” - Shaparenko, Caruana, et al. - 2005 |

20 | Ranking community answers by modeling question-answer relationships via analogical reasoning - Wang, Tu, et al. - 2009 |

19 | Overview of the 2003 kdd cup
- Gehrke, Ginsparg, et al.
(Show Context)
Citation Context ...52 63 73 81 88 101 146 182 223 59 Citation Graph The Citation dataset was crawled from the e-print arXiv that contains 421,578 citation links among a collection of 34,546 papers in the hep-ph archive =-=[33, 34]-=-. If paper i cites paper j or vice versa, then there is an undirected edge between vertex i and vertex j in the corresponding graph. This dataset was originally released in the KDD Cup 2003 [33], and ... |

19 | k-core decomposition: A tool for the analysis of large scale Internet graphs.” http://arxiv.org/ abs/cs.NI/0511007 - Alvarez-Hamelin, Dall’Asta, et al. |

19 |
Parallel community detection on large networks with propinquity dynamics, in:
- Zhang, Wang, et al.
- 2009
(Show Context)
Citation Context ...sed to improve the accuracy of community detection in different scenarios [67–70]. New measures have also been proposed to better evaluate the quality of community [71, 72]. For example, Zhang et al. =-=[67]-=- proposed a novel community detection algorithm that employs a dynamic process by contradicting network topology and topologybased propinquity. Maiya and Berger-Wolf [69] utilized a novel method based... |

16 | The web of topics: discovering the topology of topic evolution in a corpus
- Jo, Hopcroft, et al.
(Show Context)
Citation Context ...een different topics that change over time. Information-seeking activities often require the ability to identify topics with their time of appearance and to track their evolution. Recently, Jo et al. =-=[77]-=- have developed a unique approach to achieving this goal in a time-stamped 127 document collection with an underlying document network which represents a wide range of digital texts available over the... |

15 | A game-theoretic framework to identify overlapping communities
- Chen, Liu, et al.
- 2010
(Show Context)
Citation Context ...63–66]. Other models have been proposed to improve the accuracy of community detection in different scenarios [67–70]. New measures have also been proposed to better evaluate the quality of community =-=[71, 72]-=-. For example, Zhang et al. [67] proposed a novel community detection algorithm that employs a dynamic process by contradicting network topology and topologybased propinquity. Maiya and Berger-Wolf [6... |

15 | A bayesian approach toward finding communities and their evolutions in dynamic social networks
- Yang, Chi, et al.
- 2009
(Show Context)
Citation Context ...mic process by contradicting network topology and topologybased propinquity. Maiya and Berger-Wolf [69] utilized a novel method based on expander graphs to sample communities in networks. Yang et al. =-=[73]-=- explored a dynamic stochastic block model for finding communities and their evolution in dynamic social networks. However, most existing work on community detection has not considered the existence o... |

14 |
CVXOPT: A Python package for convex optimization,
- Dahl, Vandenberghe
- 2009
(Show Context)
Citation Context ... optimal value 0 to simplify the objective function L ({π1g , · · · , πMg };A). Any convex optimization package can be used to solve the optimization problem. However, regular packages such as CVXOPT =-=[50]-=- could not handle the scale of our Twitter dataset and ran out of memory. Thus, we use the limitedmemory BFGS algorithm with box constraints (L-BFGS-B) [51] to solve Eq. (4.9) and Eq. (4.10) by implic... |

12 | Collective inference on Markov models for modeling bird migration
- Sheldon, Elmohamed, et al.
- 2008
(Show Context)
Citation Context ...he sequential data (i.e., observations). The classic single path problem, solved by the Viterbi algorithm, is to find the most probable sample path given certain observations for a given Markov model =-=[81]-=-. Two generalizations of the single path problem for performing collective inference on Markov models are introduced in [81], motivated by an effort to 129 model bird migration patterns using a large ... |

10 | Detecting the structure of social networks using (α, β)-communities. Algorithms and Models for the Web Graph
- He, Hopcroft, et al.
- 2011
(Show Context)
Citation Context ...blem of community detection has been extensively studied and many algorithms have been proposed, such as cut- and conductance-based methods [11–14], spectral clustering [4, 15, 16], (α, β)-clustering =-=[10, 17]-=-, and topic modeling methods [18]. The cut- and conductance-based and spectral clustering methods are usually based on a fundamental assumption that communities have dense internal connections and spa... |

8 | Detecting community kernels in large social networks. - Wang, Lou, et al. - 2011 |

8 | J.E.: Feature-enhanced probabilistic models for diffusion network inference. - Wang, Ermon, et al. - 2012 |

8 |
Finding strongly-knit clusters in social networks.
- Mishra, Schreiber, et al.
- 2009
(Show Context)
Citation Context ...l structure, rather than due to high-degree vertices or a particular degree distribution. Detailed Introduction. We give a definition of (α, β)-community slightly different from that of Mishra et al. =-=[10]-=-. Without fixing the values of α and β, our definition highlights the contrast of internal and external connectivity. We develop a heuristic algorithm based on (α, β)-community that in practice effici... |

8 | parameterfree community discovery.
- Papadimitriou, Sun, et al.
- 2008
(Show Context)
Citation Context ... graph partitioning algorithm based on personalized PageRank vectors. An information-theoretic framework was also established to obtain an optimal partition and to find communities at multiple levels =-=[14, 58]-=-. However, communities can overlap and may also have dense external connections. Mishra et al. [10] proposed the concept of (α, β)-community and algorithms to efficiently find such communities. Ahn et... |

7 |
How does the sampling strategy impact the discovery of information diffusion in social media
- Choudhury, Lin, et al.
(Show Context)
Citation Context .... 2.2.3 Experimental Results In this section, we conduct experiments on a number of social and random graphs to demonstrate, explore, and analyze the core structure. Twitter Graph The Twitter dataset =-=[31]-=- was crawled in 2009 from the online social networking and microblogging service Twitter.com that contains friendship links among a group of Twitter users. Each vertex represents a Twitter user accoun... |

4 |
Ruby-throated Hummingbird (Archilochus colubris
- Robinson, Sargent, et al.
- 1996
(Show Context)
Citation Context ...hey generally stay above land. This prediction has been confirmed by work performed by ornithologists. For example, in the summary paragraph on migration from the Archilochus colubris species account =-=[82]-=-, Robinson et al. write “Many fly across Gulf of Mexico, but many also follow coastal route. Routes may differ for north- and south-bound birds.” The inferred distributions and paths are consistent wi... |

3 | Characterization of graphs using degree cores. Algorithms and Models for the Web Graph - Healy, Janssen, et al. - 2008 |

2 | Extracting the core structure of social networks using (α, β)-community. Internet Mathematics - Wang, Hopcroft, et al. - 2012 |

2 | Community structure in large complex networks - Wang, Hopcroft - 2010 |

1 |
Local graph partitioning using PageRank vectors
- Lang
- 2006
(Show Context)
Citation Context ... Detailed Introduction. The problem of community detection has been extensively studied and many algorithms have been proposed, such as cut- and conductance-based methods [11–14], spectral clustering =-=[4, 15, 16]-=-, (α, β)-clustering [10, 17], and topic modeling methods [18]. The cut- and conductance-based and spectral clustering methods are usually based on a fundamental assumption that communities have dense ... |

1 |
Network analysis: Methodological foundations
- Clustering
- 2005
(Show Context)
Citation Context ...re extracted and interpreted as communities by the conductance measure, which, out of numerous density-based measures, has been extensively used for detecting communities and evaluating their quality =-=[21, 22, 24]-=-. However, as clarified in Section 2.1, this type of community neither corresponds to our intuitive notion of community nor widely exist in real-world societies, where it is a matter of common observa... |

1 |
Manual of political eocnomy
- Pareto
- 1927
(Show Context)
Citation Context ...ng to multiple communities constitute the unstable regions of the network. 64 CHAPTER 3 HOW TO IDENTIFY DIFFERENT LEVELS OF SOCIAL INFLUENCE? 3.1 Introduction The Pareto principle (a.k.a. 80-20 rule) =-=[35]-=- exists almost everywhere. For example, 80% of a country’s land is owned by 20% of the population, and 80% of a company’s sales revenue comes from 20% of its clients. This is also the case for many so... |