Efficient algorithms for game-theoretic betweenness centrality

Group centralization of network indices

In network analysis, the betweenness centrality of a node informally captures\nthe fraction of shortest paths visiting that node. The computation of the\nbetweenness centrality measure is a fundamental task in the analysis of modern\nnetworks, enabling the identification of the most central nodes in such\nnetworks. Additionally to being massive, modern networks also contain\ninformation about the time at which their events occur. Such networks are often\ncalled temporal networks. The temporal information makes the study of the\nbetweenness centrality in temporal networks (i.e., temporal betweenness\ncentrality) much more challenging than in static networks (i.e., networks\nwithout temporal information). Moreover, the exact computation of the temporal\nbetweenness centrality is often impractical on even moderately-sized networks,\ngiven its extremely high computational cost. A natural approach to reduce such\ncomputational cost is to obtain high-quality estimates of the exact values of\nthe temporal betweenness centrality. In this work we present ONBRA, the first\nsampling-based approximation algorithm for estimating the temporal betweenness\ncentrality values of the nodes in a temporal network, providing rigorous\nprobabilistic guarantees on the quality of its output. ONBRA is able to compute\nthe estimates of the temporal betweenness centrality values under two different\noptimality criteria for the shortest paths of the temporal network. In\naddition, ONBRA outputs high-quality estimates with sharp theoretical\nguarantees leveraging on the \\emph{empirical Bernstein bound}, an advanced\nconcentration inequality. Finally, our experimental evaluation shows that ONBRA\nsignificantly reduces the computational resources required by the exact\ncomputation of the temporal betweenness centrality on several real world\nnetworks, while reporting high-quality estimates with rigorous guarantees.\n

Epidemics of certain viruses in a population can have major impact effects, as is the case in the recent global pandemic caused by the COVID-19 virus. Identifying infected individuals during the course of an epidemic is extremely important for measuring spread and designing more effective control measures. However, in some epidemics infected individuals do not exhibit clear symptoms despite being infected and contributing to the contagion of others (called asymptomatic). This work addresses the problem of identifying asymptomatic individuals in network epidemics based on the observation of infected (symptomatic) individuals. The main contribution of this work is the evaluation of different centrality measures to identify asymptomatic individuals when a fraction of the infected nodes in a network epidemic is observed at a given moment in time. In particular, a variation of the betweenness centrality measure is proposed in this work. An evaluation using different network models and different asymptomatic rates shows that the proposed centrality measure outperforms other centrality measures in many scenarios. Furthermore, the performance of centrality measures increases as the fraction of asymptomatic decreases, showing an interesting trade-off.

Due to serious demand-supply imbalances, many airports around the world are highly congested. Access to these highly congested (Level 3, coordinated) airports is controlled through the use of the IATA World Airport Slot allocation Guidelines (WASG). At an individual airport, slots requested by each airline are allocated at the airport under consideration independently without taking into account the interactions between slots allocated at different airports. However, in order for the air-transport network to operate seamlessly, ensuring network-wide connectivity of flights, and the interdependencies existing between the slots allocated at individual airports need to be considered. Several models have been proposed in the literature to deal with the optimum allocation of slots at a single airport. However, the literature currently does not adequately address the network-wide slot allocation problem. In this paper, we are introducing a novel approach to address the network-wide slot allocation problem. Our approach considers as an input the individual airport schedules generated during the slot allocation process at individual airports and optimally adjusts them to ensure network-wide flight connectivity by taking into account the interdependencies existing between flights connecting pairs of airports.To this end, we propose bi-objective mathematical models, which consider schedule efficiency and inter-airline fairness objectives, and incorporate the importance that different airports have for the functioning of the air transport network, using the IATA connectivity indices and the betweenness centrality measures. We solve the proposed models using the ε − constraint method to investigate trade-offs between network-wide schedule efficiency and fairness, and we investigate the effect of these trade-offs on the airlines and the airports. Results from the application of the proposed models to a test network suggest that the consideration of the contribution of the airports to network connectivity affect the way that the total network-wide schedule displacement is distributed among the airports. Specifically, we found that the use of the IATA connectivity index tends to allocate less schedule displacement to airports with frequent flights to many destinations, while the use of the betweenness centrality measure allocates less schedule displacement to airports that are more critical in ensuring the connectivity of other airports in the network.

Data classification is a major machine learning paradigm, which has been widely applied to solve a large number of real-world problems. Traditional data classification techniques consider only physical features (e.g., distance, similarity, or distribution) of the input data. For this reason, those are called \textit{low-level} classification. On the other hand, the human (animal) brain performs both low and high orders of learning and it has a facility in identifying patterns according to the semantic meaning of the input data. Data classification that considers not only physical attributes but also the pattern formation is referred to as \textit{high-level} classification. Several high-level classification techniques have been developed, which make use of complex networks to characterize data patterns and have obtained promising results. In this paper, we propose a pure network-based high-level classification technique that uses the betweenness centrality measure. We test this model in nine different real datasets and compare it with other nine traditional and well-known classification models. The results show us a competent classification performance.

오피니언 마이닝과 네트워크 분석을 활용한 상품 커뮤니티 분석: 영화 흥행성과 예측 사례

This paper is concerned with distributed computation of several commonly used centrality measures in complex networks. In particular, we propose deterministic algorithms, which converge in finite time, for the distributed computation of the degree, closeness and betweenness centrality measures in directed graphs. Regarding eigenvector centrality, we consider the PageRank problem as its typical variant, and design distributed randomized algorithms to compute PageRank for both fixed and time-varying graphs. A key feature of the proposed algorithms is that they do not require to know the network size, which can be simultaneously estimated at every node, and that they are clock-free. To address the PageRank problem of time-varying graphs, we introduce the novel concept of persistent graph, which eliminates the effect of spamming nodes. Moreover, we prove that these algorithms converge almost surely and in the sense of $L^p$. Finally, the effectiveness of the proposed algorithms is illustrated via extensive simulations using a classical benchmark.

Decision making is a critical challenge under disruptive conditions. Information flow as critical component to decision making may exacerbate or mitigate the impact of disruptive conditions. Such highly complex and unpredictable conditions produce a high demand for information. Extant literature was found to be critical of unidirectional and sender-oriented models in crisis communication which essentially blocks the feedback loop and paralyzes the crisis managers to roll-out responses. However, the pervasive social media has created new patterns of dynamic information exchange to empower the individuals to be engaged in a more evidence-based participatory form of crisis communication to frame decisions. This study used social media analytics to identify the decision clusters arising out of information exchange over social media under disruptive uncertainties. The Theory of Complexity by Edgar Morin was used as the epistemological foundation of the study. Crawled data was used from the Tourism Tribe Facebook fan page. Normalized degree and betweenness centrality measures were used to analyse the semantic networks. Tourism Tribe exhibited the highest degree centralities for both before and during the pandemic-driven disruptions. Four individuals (names changed for anonymity) were identified to be the most influential in decision making based on information shared. The causality between uncertainty and decision revealed a wide array of decision parameters – health precautions to alternative leisure engagement. The study provided practical implications for a number of stakeholders. The tourism service providers and product designers are likely to get deeper insights into the choice and priorities of the tourists confronting disruptive conditions and may re-strategise tourism offers. The alternative leisure and recreational platforms (OTT, gaming platforms etc.) may introduce new products and/or re-design their schemes.

Roadside Units (RSUs) play a vital role in 5G-enabled vehicular networks as a platform for deploying various system components such as sensing devices, access points, SDN nodes, edge/fog servers, and even light-weight cloud terminals. Therefore, it is important to place RSUs at appropriate locations such that vehicles can make frequent contacts with them and, thereby, achieving the required QoS performance. In this paper, we define a new metric Effort for placing RSUs in vehicular networks for improving the number of vehicle-RSU contacts by considering road network characteristics such as road-length, road-type, and importance of the junctions. We derive an Effort-graph from the road network to compute the RSU locations using betweenness centrality measure. The simulation results show that our scheme provides better vehicle-RSU contacts in real-world traffic scenarios without compromising vehicle-RSU contact time and per-trip contact probability.

Named Data Networking is a new networking architecture proposed as a potential design for the next generation of networks. However, supporting the mobility of producers in NDN-based IoT systems remains an open challenge for researchers to address. To tackle this problem, the NDN research group introduced a solution called KITE [1]. Since this solution is an initial design, some issues to be resolved remain, including the rendezvous placement selection. The location of the rendezvous servers can significantly reduce the path stretch efficiently. This paper focuses on improving the KITE solution by addressing two specific issues. First, we aim to determine the optimal number of rendezvous servers to use in the network. Second, we seek to identify the best locations for the rendezvous servers to minimize the path stretch efficiently. We propose an evaluation metric called average hop count based on network centrality measures to select the number and the location for the rendezvous. We evaluate the Average hop count under different network centrality measures using random scale-free network topology. Our results show that the betweenness centrality measure performs better than other measures, and we, therefore, suggest using it as a metric for selecting the number of rendezvous servers and their locations.

Proteins and the networks they determine, called interactome networks, have received attention at an important degree during the last years, because they have been discovered to have an influence on some complex biological phenomena, such as problematic disorders like cancer. This paper presents a new parallel computation technique that allows for an accurate and fast analysis of the human interactome network to be conducted. It constitutes, essentially, a proteomic data analysis process that takes into consideration the sparse nature of interactome networks. Thus, the first stage of the analysis involves the parallel computation of each proteins betweenness centrality measure through a parallel sparse networks-dedicated approach. Then, the second phase detects the functionally-related communities of proteins. In order to accomplish this purpose, we make use of a community detection algorithm that is based on the edge betweenness calculation and that has been already described in one of the authors previous papers. The new protein data analysis technique was carefully tested on real biological data and the results acknowledge the existence of some important properties of those proteins that participate in the carcinogenesis process. Apart from being particularly useful for research purposes, the novel technique also speeds up the proteomic databases analysis process, as compared to our own sequential approach. The results of the comprehensive battery of tests that were applied prove some unique topological features of cancer mutated proteins, and a possible optimization solution for cancer drugs design is suggested.

In the discussion concerning network interactions above I gave several example of how to use network theory to intervene against the spread of infectious diseases. In general a policy maker would like to lengthen the characteristic path length in the network. In doing so she may intervene by vaccinating individuals with high degree or high betweenness centrality measures. In doing so it makes the structure of the interaction network less able to efficiently spread an infectious disease. There have been many studies that have looked at these types of strategies in real world populations. For instance, I, along with Philip Polgreen and Alberto Segre, look at how allocating a limited amount of vaccine by giving it first to the individuals with highest degree in the network, as opposed to randomly, greatly limits the spread of infectious disease in a hospital. Others use social interaction data to investigate how different vaccination programs change the expected size of epidemics.

Regulatory authorities of e.g. the automotive, avionic, or railway domain pose high requirements concerning safety assurance activities. One activity known from requirements engineering (RE) is requirements traceability. We propose a model-based methodology to facilitate trace link visualization over several levels of abstraction and different viewpoints. This intuitive visualization concept allows to see (i) the artifact importance using the degree and betweenness centrality measure, and (ii) the impact of an artifact change to other artifacts. Based on this, a novel impact analysis algorithm is proposed.

We discuss how the topology of the spring system/network affects its ability to learn a desired mechanical behaviour. To ensure such a behaviour, physical parameters of springs of the system are adjusted by an appropriate gradient descent learning algorithm. We find the betweenness centrality measure particularly convenient to describe topology of the spring system structure with the best mechanical properties. We apply our results to refine an algorithm generating the structure of a spring network. We also present numerical results confirming our statements.KeywordsSpring systemMechanical behaviour problemBetweenness centrality measure

Application of the load flow and random flow models for the analysis of power transmission networks