Methods to find community based on edge centrality

Divisive algorithms are of great importance for community detection in complex networks. One algorithm proposed by Girvan and Newman (GN) based on an edge centrality named betweenness, is a typical representative of this field. Here we studied three edge centralities based on network topology, walks and paths respectively to quantify the relevance of each edge in a network, and proposed a divisive algorithm based on the rationale of GN algorithm for finding communities that removes edges iteratively according to the edge centrality values in a certain order. In addition, we gave a comparison analysis of these measures with the edge betweenness and information centrality. We found the principal difference among these measures in the partition procedure is that the edge centrality based on walks first removes the edge connected with a leaf vertex, but the others first delete the edge as a bridge between communities. It indicates that the edge centrality based on walks is harder to uncover communities than other edge centralities. We also tested these measures for community detection. The results showed that the edge information centrality outperforms other measures, the edge centrality based on walks obtains the worst results, and the edge betweenness gains better performance than the edge centrality based on network topology. We also discussed our method’s efficiency and found that the edge centrality based on walks has a high time complexity and is not suitable for large networks.     

Betweenness centrality in finite components of complex networks

We use generating function formalism to obtain an exact formula of the betweenness centrality in finite components of random networks with arbitrary degree distributions. The formula is obtained as a function of the degree and the component size, and is confirmed by simulations for Poisson, exponential, and power-law degree distributions. We find that the betweenness centralities for the three distributions are asymptotically power laws with an exponent 1.5 and are invariant to the particular distribution parameters.     

节点中心性对复杂网络传播模式的影响分析

在众多的重要节点评估方法研究中,具有较高中心性的节点一直是关注的焦点,许多传播行为的研究也主要围绕高中心性节点展开,因此在一定程度上忽略了低中心性节点对传播行为的影响.本文从传播异构性角度,通过初始感染最大中心性节点和最小中心性节点揭示网络结构异构性对信息传播的影响.实验结果表明,传播过程中存在"链型"和"扇型"两种传播模式,在初始感染比例不断提升的情况下,两种传播模式的相互转换引发传播速率的变化,进一步促使非线性传播规模交叉现象的产生.这一现象说明,在宏观的信息传播过程中,最小中心性节点的影响力不容忽视,尤其在初始感染比例升高时,最小中心性节点比最大中心性节点更具传播优势.     

The importance of centralities in dark network value chains

This paper introduces three novel centrality measures based on the nodes’ role in the operation of a joint task, i.e., their position in a criminal network value chain. For this, we consider networks where nodes have attributes describing their “capabilities” or “colors”, i.e., the possible roles they may play in a value chain. A value chain here is understood as a series of tasks to be performed in a specific order, each requiring a specific capability. The first centrality notion measures how many value chain instances a given node participates in. The other two assess the costs of replacing a node in the value chain in case the given node is no longer available to perform the task. The first of them considers the direct distance (shortest path length) between the node in question and its nearest replacement, while the second evaluates the actual replacement process, assuming that preceding and following nodes in the network should each be able to find and contact the replacement. In this report, we demonstrate the properties of the new centrality measures using a few toy examples and compare them to classic centralities, such as betweenness, closeness and degree centrality. We also apply the new measures to randomly colored empirical networks. We find that the newly introduced centralities differ sufficiently from the classic measures, pointing towards different aspects of the network. Our results also pinpoint the difference between having a replacement node in the network and being able to find one. This is the reason why “introduction distance” often has a noticeable correlation with betweenness. Our studies show that projecting value chains over networks may significantly alter the nodes’ perceived importance. These insights might have important implications for the way law enforcement or intelligence agencies look at the effectiveness of dark network disruption strategies over time.     

Random walks on complex networks

We investigate random walks on complex networks and derive an exact expression for the mean first-passage 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.     

Error and attack tolerance of synchronization in Hindmarsh–Rose neural networks with community structure

Synchronization is one of the most important features observed in large-scale complex networks of interacting dynamical systems. As is well known, there is a close relation between the network topology and the network synchronizability. Using the coupled Hindmarsh–Rose neurons with community structure as a model network, in this paper we explore how failures of the nodes due to random errors or intentional attacks affect the synchronizability of community networks. The intentional attacks are realized by removing a fraction of the nodes with high values in some centrality measure such as the centralities of degree, eigenvector, betweenness and closeness. According to the master stability function method, we employ the algebraic connectivity of the considered community network as an indicator to examine the network synchronizability. Numerical evidences show that the node failure strategy based on the betweenness centrality has the most influence on the synchronizability of community networks. With this node failure strategy for a given network with a fixed number of communities, we find that the larger the degree of communities, the worse the network synchronizability; however, for a given network with a fixed degree of communities, we observe that the more the number of communities, the better the network synchronizability.     

Walks on Weighted Networks

We investigate the dynamics of random walks on weighted networks. Assuming that the edge weight and the node strength are used as local information by a random walker. Two kinds of walks, weight-dependent walk and strength-dependent walk, are studied. Exact expressions for stationary distribution and average return time are derived and confirmed by computer simulations. The distribution of average return time and the mean-square displacement are calculated for two walks on the Barrat-Barthelemy-Vespignani （BBV） networks. It is found that a weight-dependent walker can arrive at a new territory more easily than a strength-dependent one.     

Walks on Apollonian networks

We carry out comparative studies of random walks on deterministic Apollonian networks (DANs) and random Apollonian networks (RANs). We perform computer simulations for the mean first-passage time, the average return time, the mean-square displacement, and the network coverage for the unrestricted random walk. The diffusions both on DANs and RANs are proved to be sublinear. The effects of the network structure on the dynamics and the search efficiencies of walks with various strategies are also discussed. Contrary to intuition, it is shown that the self-avoiding random walk, which has been verified as an optimal local search strategy in networks, is not the best strategy for the DANs in the large size limit.     

Optimal navigation for characterizing the role of the nodes in complex networks

In this paper, we explore how the approach of optimal navigation (Cajueiro (2009) [33]) can be used to evaluate the centrality of a node and to characterize its role in a network. Using the subway network of Boston and the London rapid transit rail as proxies for complex networks, we show that the centrality measures inherited from the approach of optimal navigation may be considered if one desires to evaluate the centrality of the nodes using other pieces of information beyond the geometric properties of the network. Furthermore, evaluating the correlations between these inherited measures and classical measures of centralities such as the degree of a node and the characteristic path length of a node, we have found two classes of results. While for the London rapid transit rail, these inherited measures can be easily explained by these classical measures of centrality, for the Boston underground transportation system we have found nontrivial results.     

Measuring the flow of information among cities using the diffusion power

In this Letter, we define here the so-called diffusion power - an extension of the dominance power, which considers the interaction between neighbors of higher orders. Using this measure, we analyze the centrality of cities in two networks of the flow of information among these cities, namely a network of calls among the cities and a network of radio stations. Finally, we explain the centralities of the cities evaluated using the diffusion power in terms of the specific characteristics of the cities that belong to the network.     

Biased trapping issue on weighted hierarchical networks

In this paper, we present trapping issues of weight-dependent walks on weighted hierarchical networks which are based on the classic scale-free hierarchical networks. Assuming that edge’s weight is used as local information by a random walker, we introduce a biased walk. The biased walk is that a walker, at each step, chooses one of its neighbours with a probability proportional to the weight of the edge. We focus on a particular case with the immobile trap positioned at the hub node which has the largest degree in the weighted hierarchical networks. Using a method based on generating functions, we determine explicitly the mean first-passage time (MFPT) for the trapping issue. Let parameter a (0 < a < 1) be the weight factor. We show that the efficiency of the trapping process depends on the parameter a; the smaller the value of a, the more efficient is the trapping process.     

Computing Influential Nodes Using the Nearest Neighborhood Trust Value and PageRank in Complex Networks

Computing influential nodes gets a lot of attention from many researchers for information spreading in complex networks. It has vast applications, such as viral marketing, social leader creation, rumor control, and opinion monitoring. The information-spreading ability of influential nodes is greater compared with other nodes in the network. Several researchers proposed centrality measures to compute the influential nodes in a complex network, such as degree, betweenness, closeness, semi-local centralities, and PageRank. These centrality methods are defined based on the local and/or global information of nodes in the network. However, due to their high time complexity, centrality measures based on the global information of nodes have become unsuitable for large-scale networks. Very few centrality measures exist that are based on the attributes between nodes and the structure of the network. We propose the nearest neighborhood trust PageRank (NTPR) based on the structural attributes of neighbors and nearest neighbors of nodes. We define the measure based on the degree ratio, the similarity between nodes, the trust values of neighbors, and the nearest neighbors. We computed the influential nodes in various real-world networks using the proposed centrality method. We found the maximum influence by using influential nodes with SIR and independent cascade methods. We also compare the maximum influence of our centrality measure with the existing basic centrality measures.     

Community detection using global and local structural information

Community detection is of considerable importance for understanding both the structure and function of complex networks. In this paper, we introduced the general procedure of the community detection algorithms using global and local structural information, where the edge betweenness and the local similarity measures respectively based on local random walk dynamics and local cyclic structures were used. The algorithms were tested on artificial and real-world networks. The results clearly show that all the algorithms have excellent performance in the tests and the local similarity measure based on local random walk dynamics is superior to that based on local cyclic structures.     

Two Novel Methods to Enhance Network Synchronizability

In this paper, we propose two methods to enhance the synchronizability of a class of complex networks which do not hold the positive correlation between betweenness centrality （BC） and degree of a node, and observe other topology characteristics of the network affected by the methods. Numerical simulations show that both methods can effectively enhance the synchronizability of this kind of networks. Furthermore, we show that the maximal BC of all edges is an important factor to affect the network synchronizability, although it is not the unique factor.     

Lobby index as a network centrality measure

We study the lobby index (l$l$-index for short) as a local node centrality measure for complex networks. The l$l$-index is compared with degree (a local measure), betweenness and Eigenvector centralities (two global measures) in the case of a biological network (Yeast interaction protein–protein network) and a linguistic network (Moby Thesaurus   II). In both networks, the l$l$-index has a poor correlation with betweenness but correlates with degree and Eigenvector centralities. Although being local, the l$l$-index carries more information about its neighbors than degree centrality. Also, it requires much less time to compute when compared with Eigenvector centrality. Results show that the l$l$-index produces better results than degree and Eigenvector centrality for ranking purposes.     

Entropy and Network Centralities as Intelligent Tools for the Investigation of Terrorist Organizations

In recent years, law enforcement authorities have increasingly used mathematical tools to support criminal investigations, such as those related to terrorism. In this work, two relevant questions are discussed: “How can the different roles of members of a terrorist organization be recognized?” and “are there early signs of impending terrorist acts?” These questions are addressed using the tools of entropy and network theory, more specifically centralities (degree, betweenness, clustering) and their entropies. These tools were applied to data (physical contacts) of four real terrorist networks from different countries. The different roles of the members are clearly recognized from the values of the selected centralities. An early sign of impending terrorist acts is the evolutionary pattern of the values of the entropies of the selected centralities. These results have been confirmed in all four terrorist networks. The conclusion is expected to be useful to law enforcement authorities to identify the roles of the members of terrorist organizations as the members with high centrality and to anticipate when a terrorist attack is imminent, by observing the evolution of the entropies of the centralities.     

Finding communities in directed networks by PageRank random walk induced network embedding

Community structure has been found to exist ubiquitously in many different kinds of real world complex networks. Most of the previous literature ignores edge directions and applies methods designed for community finding in undirected networks to find communities. Here, we address the problem of finding communities in directed networks. Our proposed method uses PageRank random walk induced network embedding to transform a directed network into an undirected one, where the information on edge directions is effectively incorporated into the edge weights. Starting from this new undirected weighted network, previously developed methods for undirected network community finding can be used without any modification. Moreover, our method improves on recent work in terms of community definition and meaning. We provide two simulated examples, a real social network and different sets of power law benchmark networks, to illustrate how our method can correctly detect communities in directed networks.     

Rethinking failure and attack tolerance assessment in complex networks

Studies have revealed that real complex networks are inherently vulnerable to the loss of high centrality nodes. These nodes are crucial to maintaining the network connectivity and are identified by classical measures, such as degree and betweenness centralities. Despite its significance, an assessment based solely on this vulnerability premise is misleading for the interpretation of the real state of the network concerning connectivity. As a matter of fact, some networks may be in a state of imminent fragmentation before such a condition is fully characterized by an analysis targeted solely on the centrally positioned nodes. This work aims at showing that, in fact, it is basically the global network configuration that is responsible for network fragmentation, as it may allow many other lower centrality nodes to seriously damage the network connectivity.     

The backbone of a city

Recent studies have revealed the importance of centrality measures to analyze various spatial factors affecting human life in cities. Here we show how it is possible to extract the backbone of a city by deriving spanning trees based on edge betweenness and edge information. By using as sample cases the cities of Bologna and San Francisco, we show how the obtained trees are radically different from those based on edge lengths, and allow an extended comprehension of the “skeleton” of most important routes that so much affects pedestrian/vehicular flows, retail commerce vitality, land-use separation, urban crime and collective dynamical behaviours.     

网络集聚性对节点中心性指标的准确性影响

节点中心性指标是从特定角度对网络某一方面的结构特点进行刻画的度量指标, 因此网络拓扑结构的改变会对节点中心性指标的准确性产生重要影响. 本文利用Holme-Kim模型构建可变集聚系数的无标度网络, 然后采用Susceptible-Infective-Removal模型进行传播影响力的仿真实验, 接着分析了节点中心性指标在不同集聚系数的无标度网络中的准确性. 结果表明, 度中心性和介数中心性的准确性在低集聚系数的网络中表现更好, 特征向量中心性则在高集聚类网络中更准确, 而紧密度中心性的准确性受网络集聚系数的变化影响较小. 因此当网络的集聚系数较低时, 可选择度或者介数作为中心性指标进行网络节点影响力评价; 反之则选择紧密度指标或特征向量指标较好, 尤其当网络的集聚系数接近0.6时特征向量的准确性可以高达到0.85, 是度量小规模网络的较优选择. 另一方面, 传播过程的感染率越高, 度指标和介数指标越可靠, 紧密度和特征向量则相反. 最后Autonomous System实证网络的断边重连实验, 进一步验证了网络集聚性的改变会对节点中心性指标的准确性产生重要影响.