Search

This paper investigates the stability and evolution of the world stage of global science at the city level by analyzing changes in co-authorship network centrality rankings over time. Driven by the problem that there exists no consensus in the literature on how the spatial unit “city” should be defined, we first propose a new approach to delineate so-called scientific cities. On a high-quality Web of Science dataset of 21.5 million publications over the period 2008–2020, we study changes in centrality rankings of subsequent 3-year time-slices of scientific city co-authorship networks at various levels of impact. We find that, over the years, the world stage of global science has become more stable. Additionally, by means of a comparison with degree respecting rewired networks we reveal how new co-authorships between authors from previously unconnected cities more often connect ‘close’ cities in the network periphery.

The goal of this paper is to learn the dynamics of truck co-driving behaviour. Understanding this behaviour is important because co-driving has a potential positive impact on the environment. In the so-called co-driving network, trucks are nodes while links indicate that two trucks frequently drive together. To understand the network’s dynamics, we use a link prediction approach employing a machine learning classifier. The features of the classifier can be categorized into spatio-temporal features, neighbourhood features, path features, and node features. The very different types of features allow us to understand the social processes underlying the co-driving behaviour. Our work is based on a spatio-temporal data not studied before. Data is collected from 18 million truck movements in the Netherlands. We find that co-driving behaviour is best described by using neighbourhood features, and to lesser extent by path and spatio-temporal features. Node features are deemed unimportant. Findings suggest that the dynamics of a truck co-driving network has clear social network effects. © 2020, Springer Nature Switzerland AG.

In link prediction, the goal is to predict which links will appear in the future of an evolving network. To estimate the performance of these models in a supervised machine learning model, disjoint and independent train and test sets are needed. However, objects in a real-world network are inherently related to each other. Therefore, it is far from trivial to separate candidate links into these disjoint sets.Here we characterize and empirically investigate the two dominant approaches from the literature for creating separate train and test sets in link prediction, referred to as random and temporal splits. Comparing the performance of these two approaches on several large temporal network datasets, we find evidence that random splits may result in too optimistic results, whereas a temporal split may give a more fair and realistic indication of performance. Results appear robust to the selection of temporal intervals. These findings will be of interest to researchers that employ link prediction or other machine learning tasks in networks.

The main topic of this paper is the discovery of motifs in multiplex corporate networks. Network motifs are small subgraphs occurring at significantly higher numbers than in similar random networks. They can be seen as the building blocks of a complex network. In real-world network data, multiple types of (possibly overlapping) relationships may be present among the nodes, forming so-called multiplex networks. Detecting motifs in such networks is difficult, as existing subgraph enumeration algorithms are not directly applicable to multiplex network data. In addition, the selection of a proper multiplex null model to test the significance of the enumerated subgraphs is nontrivial. This paper addresses these two problems, resulting in three contributions. First, we present a method based on layer encoding for adequately handling the multiplex aspect in subgraph enumeration. Second, a null model is proposed that is able to preserve the relationship between the different types of links, taking into account that a particular link type may be the result of a projection from a bipartite network. Finally, we perform experiments on corporate network data from Germany, in which around 75000 75000 nodes represent corporations and roughly 195000 195000 links represent connectedness of firms based on shared board members and ownership. We demonstrate how incorporating the multiplex aspect in motif detection is able to reveal new insights that could not be obtained by studying only one type of relationship. Furthermore, results uncover how the financial sector is over-represented in the more complex motifs, hinting at a surprisingly prominent role of the financial sector in the largely industry-oriented corporate network of Germany.

This paper studies online child exploitation networks in which users communicate about illegal child pornography material. Law enforcement agencies are extremely interested in better understanding these networks and their key players. We utilize unique real-world network data sets collected from two different online discussion forums on the dark net. Our study of the network structure underlying these forums results in three contributions.First, we propose an approach to identify key players using various centrality measures, allowing us to automatically rank users. Experiments show that our method closely resembles a network-agnostic ranking of users created by domain experts. Second, network metrics are able to characterize a large portion of the users, allowing us to distinguish between regular users, managers and technical moderators. Finally, analyzing the structural properties and distributions of these networks in both the one-mode and two-mode perspective reveals various interesting network-driven insights, such as anti-lurker and anti-law enforcement policies and new user application guidelines. In addition, we found that active users form an elite that participate in more specialized discussions.

This paper proposes a novel approach to count temporal motifs in multilayer complex networks. Network motifs, i.e., small characteristic patterns of a handful of nodes and edges, have repeatedly been shown to be instrumental in understanding real-world complex systems. However, exhaustively enumerating these motifs is computationally infeasible for larger networks. Therefore, the focus of this work is on algorithms that efficiently count network motifs. This facilitates the discovery of motifs in networks with millions of nodes and edges, enabling the following three contributions of this paper. First, we propose an extension of an existing counting algorithm to also efficiently count multilayer temporal motifs. In addition to dealing with the timestamp at which a link (re-)occurs, the algorithm also efficiently counts interaction patterns across different network layers. Second, we demonstrate how partial timing, a common phenomenon in real-world settings where only part of the layers are timed, can be incorporated. Third, we assess the performance of the proposed temporal multilayer counting algorithm on a number of real-world network datasets. Experiments reveal interesting insights in the heterogen eous interplay between network layers in, for example, online expert communities, showing how particular temporal motifs are characteristic for certain layers of interaction.

This paper examines the co-driving behavior of truck drivers using network analysis. From a unique spatiotemporal dataset encompassing more than 10 million measurements of trucks passing 17 different highway locations in the Netherlands, we extract a so-called co-driving network. In this network, nodes are truck drivers and edges represent pairs of trucks that are systematically driving together. The obtained co-driving network structure has various properties common to real-world networks, such as a dominant giant component and a power law degree distribution. Moreover, network distance metrics and community detection reveal that the network has a highly modular structure. We furthermore propose a method for understanding the network community structure through attribute assortativity. Results indicate that co-driving links are mostly established based on geographical aspects: truck drivers from the same country or the same region in the Netherlands are more inclined to drive together. The resulting improved understanding of co-driving behavior has important implications for society and the environment, as trucks coordinating their driving behavior together help reduce traffic congestion and optimize fuel usage.

The overwhelming amount of network data that is nowadays available, leads to an increased demand for techniques that automatically identify anomalous nodes. Examples are network intruders in physical networks or spammers spreading unwanted advertisements in online social networks. Existing methods typically identify network anomalies from a local perspective, only considering metrics related to a node and connections in its direct neighborhood. However, such methods often miss anomalies as they overlook crucial distortions of the network structure that are only visible at the macro level. To solve these problems, in this paper, the CADA algorithm is proposed, which identifies irregular nodes from a global perspective. It does so by measuring the extent to which a node connects to man y different communities while not obviously belonging to one community itself. Results on synthetic and real-world data show that the incorporation of the community aspect is of critical importance, as our algorithm significantly outperforms previously suggested techniques. In addition, it scales well to larger networks of hundreds of thousands of nodes and millions of links. Moreover, the proposed method is parameter-free, enabling the hassle-free identification of anomalies in a wide variety of application domains.

The rapid development of World Wide Web accelerates information spreading in various ways. Thanks to the emergence of multiple social platforms, some events which are not much attractive in the past can become social hot spots nowadays. In this paper, we study the information diffusion process of “IP MAN3 box office fraud”, which is widely diffused in the largest Chinese microblogging system, namely Sina Weibo, in March 2016. Based on the temporal metric we have proposed, we succeed in finding out the sources of the information, and constructing the panorama of the diffusion process. In addition, a portion of nodes that promote the diffusion are identified by using the node importance algorithms. Finally, the users with abnormal behaviors in the process of event development are identified.

An electroencephalography (EEG) coherence network is a representation of functional brain connectivity. However, typical visualizations of coherence networks do not allow an easy embedding of spatial information or suffer from visual clutter, especially for multichannel EEG coherence networks. In this paper, a new method for data-driven visualization of multichannel EEG coherence networks is proposed to avoid the drawbacks of conventional methods. This method partitions electrodes into dense groups of spatially connected regions. It not only preserves spatial relationships between regions, but also allows an analysis of the functional connectivity within and between brain regions, which could be used to explore the relationship between functional connectivity and underlying brain structures. In addition, we employ an example to illustrate the difference between the proposed method and two other data-driven methods when applied to coherence networks in older and younger adults who perform a cognitive task. The proposed method can serve as an preprocessing step before a more detailed analysis of EEG coherence networks.

Most existing sampling algorithms on graphs (i.e., network-structured data) focus on sampling from memory-resident static graphs and assume the entire graphs are always available. However, the graphs encountered in modern applications are often too large and/or too dynamic to be processed with limited memory. Furthermore, existing sampling techniques are inadequate for preserving the inherent clustering structure, which is an essential property of complex networks. To tackle these problems, we propose a new sampling algorithm that dynamically maintains a representative sample and is capable of retaining clustering structure in graph streams at any time. Performance of the proposed algorithm is evaluated through empirical experiments using real-world networks. The experimental results have shown that our proposed \textit{CPIES} algorithm can produce clustering-structure representative samples and outperforms current online sampling algorithms.

SUMMARY Objective: Amelogenesis imperfecta (AI) is a heterogenous genetic disorder that interferes with normal enamel formation in the absence of systemic disorders. The patients' main concerns are caries susceptibility, poor esthetics, and generalized sensitivity. There is a broad clinical spectrum, from discolorations to consequent enamel alterations. This case report describes the 15-year case study and the full-mouth rehabilitation of two siblings affected by a hypocalcified AI. Clinical Considerations: In these two patients, conservative care with stainless steel crowns and direct composite restorations was undertaken to restore function and esthetics and to reduce sensitivities in primary and mixed dentitions. The difficulties in monitoring resulted in severe infectious complications (dental abscess with cutaneous fistula), important dental defects, and loss of spaces with subsequent malocclusion. In the young adult dentition, they were treated by extractions, root canal therapies, and new restorations: stainless steel crowns for permanent molars, direct composite restorations (with strip crowns) for incisors and maxillary canines (to improve the crown morphology as well as to mask the discolorations and the malpositions), and adjusted composite crown molds using a thermoforming procedure for premolars and the mandibular canines. The main difficulties were rapid tooth surface loss, bonding to atypical enamel, developing dentition, long-term follow-up. Conclusion: Restoring function and esthetics in AI-affected patients is a challenge from primary to adult dentition. Early corrections are essential to avoid dental damage and for psychological benefits. This clinical report highlights the adhesive rehabilitation for anterior and premolar areas and the difficulty of patient follow-up.