Real Network Data Research Articles

The Node Vector Distance Problem in Complex Networks

We describe a problem in complex networks we call the Node Vector Distance (NVD) problem, and we survey algorithms currently able to address it. Complex networks are a useful tool to map a non-trivial set of relationships among connected entities, or nodes. An agent—e.g., a disease—can occupy multiple nodes at the same time and can spread through the edges. The node vector distance problem is to estimate the distance traveled by the agent between two moments in time. This is closely related to the Optimal Transportation Problem (OTP), which has received attention in fields such as computer vision. OTP solutions can be used to solve the node vector distance problem, but they are not the only valid approaches. Here, we examine four classes of solutions, showing their differences and similarities both on synthetic networks and real world network data. The NVD problem has a much wider applicability than computer vision, being related to problems in economics, epidemiology, viral marketing, and sociology, to cite a few. We show how solutions to the NVD problem have a wide range of applications, and we provide a roadmap to general and computationally tractable solutions. We have implemented all methods presented in this article in a publicly available open source library, which can be used for result replication.

Channel Optimization of Marketing Based on Users’ Social Network Information

Marketing in the social network environment integrates current advanced internet and information technologies. This marketing method not only broadens marketing channels and builds a network communication platform but also meets the purchase needs of customers in the entire market and shortens customer purchases. The process is also an inevitable product of the development of the times. However, when companies use social networks for product marketing, they usually face the impact of multiple realistic factors. This article takes the maximization of influence as the main idea to find seed users for product information dissemination and also considers the users’ interest preferences. The target users can influence the product, and the company should control marketing costs to obtain a larger marginal benefit. Based on this, this paper considers factors such as the scale of information diffusion, user interest preferences, and corporate budgets, takes the influence maximization model as a multiobjective optimization problem, and proposes a multiobjective maximization of influence (MOIM) model. To solve the NP-hard problem of maximizing influence, this paper uses Monte Carlo sampling to calculate high-influence users. Next, a seed user selection algorithm based on NSGA-II is proposed to optimize the above three objective functions and find the optimal solution. We use real social network data to verify the performance of models and methods. Experiments show that the proposed model can generate appropriate seed sets and can meet different purposes of information dissemination. Sensitivity analysis proves that our model is robust under different actual conditions.

Privacy-Preserving Network Analytics

Using financial networks as a backdrop, we develop a new framework for privacy-preserving network analytics. Adopting the debt and equity models of Eisenberg and Noe (2001) and Elliott et al. (2014) as proof of concept, we show how aggregate-level statistics required for stress testing and stability assessment can be derived on real network data, without any individual node revealing its private information to any third party, be it other nodes in the network, or even a central agent. Our work helps bridge the gap between the theoretical models of financial networks that assume agents have full information, and the real world, where information sharing is hindered by privacy and security concerns.

A Node Embedding-Based Influential Spreaders Identification Approach

Node embedding is a representation learning technique that maps network nodes into lower-dimensional vector space. Embedding nodes into vector space can benefit network analysis tasks, such as community detection, link prediction, and influential node identification, in both calculation and richer application scope. In this paper, we propose a two-step node embedding-based solution for the social influence maximization problem (IMP). The solution employs a revised network-embedding algorithm to map input nodes into vector space in the first step. In the second step, the solution clusters the vector space nodes into subgroups and chooses the subgroups’ centers to be the influential spreaders. The proposed approach is a simple but effective IMP solution because it takes both the social reinforcement and homophily characteristics of the social network into consideration in node embedding and seed spreaders selection operation separately. The information propagation simulation experiment of single-point contact susceptible-infected-recovered (SIR) and full-contact SIR models on six different types of real network data sets proved that the proposed social influence maximization (SIM) solution exhibits significant propagation capability.

A Hybrid Approach for Process Monitoring: Improving Data-Driven Methodologies With Dataset Size Reduction and Interval-Valued Representation

Kernel principal component analysis (KPCA) is a well-established data-driven process modeling and monitoring framework that has long been praised for its performances. However, it is still not optimal for large-scale and uncertain systems. Applying KPCA usually takes a long time and a significant storage space when big data are utilized. In addition, it leads to a serious loss of information and ignores uncertainties in the processes. Consequently, in this paper, two uncertain nonlinear statistical fault detection methods using an interval reduced kernel principal component analysis (IRKPCA) are proposed. The main objective of the proposed methods is twofold. Firstly, reduce the number of observations in the data matrix through two techniques: a method, called IRKPCA $_{ED}$ , is based on Euclidean distance between samples as dissimilarity metric such that only one observation is kept in case of redundancy to build the reduced reference KPCA model, and another method, called IRKPCA $_{PCA}$ , is established on the PCA algorithm to treat the hybrid correlations between process variables and extract a reduced number of observations from the training data matrix. Secondly, address the problem of uncertainties in systems using a latent-driven technique based on interval-valued data. Taking into account sensors uncertainties via IRKPCA ensures better monitoring by reducing the computational and storage costs. The study demonstrated the feasibility and effectiveness of the proposed approaches for faults detection in two real world applications: Tennessee Eastman (TE) process and real air quality monitoring network (AIRLOR) data.

Time-Varying Graph Signal Denoising via Median Filters

Graph Signal Processing (GSP) leverages pair-wise relationship between nodes of a graph to formulate operators on signals defined over the nodes. Most existing graph signal operators in the literature are linear, and can be described by linear transformation matrices. Recently, works are emerging that consider the time correlation of graph signals, leading to time-vertex signal processing. By exploiting the joint correlations across the graph topology and time, better results can be obtained. In this brief, we propose a median operator that will leverage the joint correlation, for denoising time-varying graph signals. The median operator, known for its robustness to outliers in statistics, has been very successful in traditional signal processing, especially for images. The efficient and highly localised graph median filters developed here are applied to denoising real world sensor network data. Real world sensor nodes are usually resource limited in terms of their computational and communication capacity, hence the imperative requirement for efficient localised filters. Comprehensive experimental results will show that in some cases, the performance is significantly better than the equivalent linear operator counterpart.

An Enhanced Network-Consistent Travel Speed Generation Scheme on Time-Dependent Shortest Path and Routing Problems

The time-dependent shortest path and vehicle routing literature depends on realistic and reasonable test data for demonstration and performance evaluation. Despite the advancements in GPS and tracking technologies there is still lack and inaccessibility of publicly available real-world road networks with time-dependent arc costs and speeds. Since most of the time-dependent travel time layer generation models proposed for vehicle routing problems (VRPs) are mainly developed for synthetic networks, they cannot capture some realistic features of the real road networks and cannot be used in time-dependent shortest path problems (TDSPPs). In this paper, we first exploit spatial and temporal behavior of travel times using real life road network and speed data, and discuss the cases where the existing methods in the literature are not applicable. Then, we propose an enhanced method that is best fitted for TDSPP and time-dependent VRP (TDVRP). The proposed method can be implemented on both synthetic and real road networks. Finally, we apply our method to generate realistic speed data on Istanbul road network and demonstrate the applicability in TDSPP and TDVRP.

Research on information steganography based on network data stream

To protect the information from being intercepted by third parties during the network communication process, this paper proposes a new type of data steganography technology based on network data flow. Using the network protocol itself and the relationship between data packets in the entire network data stream to perform network data steganography, transfer hidden data, and perform secondary identity authentication. Different from the traditional steganography method, this method can encode the hidden data and send the interval value by embedding the data packet, thereby hiding and transmitting the hidden data. In this technology, the operation of hidden data does not affect the user's access request for real network data, and it can perform processes such as hidden data transfer and secondary authentication without the user being able to detect it. Through experimental verification and evaluation, our method improves the concealment of the steganographic channel, is not easy to attract attention and has no obvious statistical characteristics of the traffic, and can improve the concealment and robustness of the steganography technology based on network data streams.

Complex network classification with convolutional neural network

Classifying large scale networks into several categories and distinguishing them according to their fine structures is of great importance with several applications in real life. However, most studies of complex networks focus on properties of a single network but seldom on classification, clustering, and comparison between different networks, in which the network is treated as a whole. Due to the non-Euclidean properties of the data, conventional methods can hardly be applied on networks directly. In this paper, we propose a novel framework of complex network classifier (CNC) by integrating network embedding and convolutional neural network to tackle the problem of network classification. By training the classifiers on synthetic complex network data and real international trade network data, we show CNC can not only classify networks in a high accuracy and robustness, it can also extract the features of the networks automatically.

Equitable optimized airspace sectorization based on constraint programming and OWA aggregation

PurposeAirspace sectorization is an important task, which has a significant impact in the everyday work of air control services. Especially in recent years, because of the constant increase in air traffic, existing airspace sectorization techniques have difficulties to tackle the large air traffic volumes, creating imbalanced sectors and uneven workload distribution among sectors. The purpose of this paper is to propose a new approach to find optimal airspace sectorization balancing the traffic controller workload between sectors, subject to airspace requirements.Design/methodology/approachA constraint programming (CP) model called equitable airspace sectorization problem (EQASP) relies on ordered weighted averaging (OWA) multiagent optimization and the parallel portfolio architecture has been developed, which integrates the equity into an existing CP approach (Trandac et al., 2005). The EQASP was evaluated and compared with the method of Trandac et al. (2005), according to the quality of workload balancing between sectors and the resolution performance. The comparison was achieved using real air traffic low-altitude network data sets of French airspace for five flight information regions for 24 h a day and the Algerian airspace for three various periods (off peak hours, peak hours and 24 h).FindingsIt has been demonstrated that the proposed EQASP model, which is based on OWA multicriteria optimization method, significantly improved both the solving performance and the workload equity between sectors, while offering strong theoretical properties of the balancing requirement. Interestingly, when solving hard instances, our parallel sectorization tool can provide, at any time, a workable solution, which satisfies all geometric constraints of sectorization.Practical implicationsThis study can be used to design well-balanced air sectors in terms of workload between control units in the strategic phase. To fulfil the airspace users’ constraints, one can refer to this study to assess the capacity of each air sector (especially the overloaded sectors) and then adjust the sector’s shape to respond to the dynamic changes in traffic patterns.Social implicationsThis theoretical and practical approach enables the development and support of the definition of the “Air traffic management (ATM) Concept Target” through improvements in human factors specifically (balancing workload across sectors), which contributes to raising the level of capacity, safety and efficiency (SESAR Vision of ATM 2035).Originality/valueIn their approach, the authors proposed an OWA-based multiagent optimization model, ensuring the search for the best equitable solution, without requiring user-defined balancing constraints, which enforce each sector to have a workload between two user-defined bounds (Wmin, Wmax).

Bayesian estimation of the latent dimension and communities in stochastic blockmodels

Spectral embedding of adjacency or Laplacian matrices of undirected graphs is a common technique for representing a network in a lower dimensional latent space, with optimal theoretical guarantees. The embedding can be used to estimate the community structure of the network, with strong consistency results in the stochastic blockmodel framework. One of the main practical limitations of standard algorithms for community detection from spectral embeddings is that the number of communities and the latent dimension of the embedding must be specified in advance. In this article, a novel Bayesian model for simultaneous and automatic selection of the appropriate dimension of the latent space and the number of blocks is proposed. Extensions to directed and bipartite graphs are discussed. The model is tested on simulated and real world network data, showing promising performance for recovering latent community structure.

Trust Relationship Prediction in Alibaba E-Commerce Platform

This paper introduces how to infer trust relationships from billion-scale networked data to benefit Alibaba E-Commerce business. To effectively leverage the network correlations between labeled and unlabeled relationships to predict trust relationships, we formalize trust into multiple types and propose a graphical model to incorporate type-based dyadic and triadic correlations, namely eTrust. We also present a fast learning algorithm in order to handle billion-scale networks. Systematically, we evaluate the proposed methods on four different genres of datasets with labeled trust relationships: Alibaba, Epinions, Ciao, and Advogato. Experimental results show that the proposed methods achieve significantly better performance than several comparison methods (+1.7-32.3% by accuracy; $p p 0 . 01 , with $t$ t -test). Most importantly, when handling the real large networked data with over 1,200,000,000 edges (Ali-large), our method achieves 2,000× speedup to infer trust relationships, comparing with the traditional graph learning algorithms. Finally, we have applied the inferred trust relationships to Alibaba E-commerce platform: Taobao, and achieved 2.75 percent improvement on gross merchandise volume (GMV).

Analysis and comparison of centrality measures applied to urban networks with data

For a considerable time, researchers have focused on defining different measures capable to characterizing the importance of vertices in networks. One type of these networks, the cities, are complex systems that generate large quantity of information. These data are an important part of the characteristics of the urban network itself. Because of this, it is crucial to have a classification system, for the vertices of a network, considering the data we can find in the city itself. To address this question, this paper studies and compares several measures of centrality specifically applied to urban networks. These centralities are based on the calculation of the eigenvectors of a matrix and are very suitable for urban networks with data. With the aim of expanding the range covered by these measures, a new centrality measure is presented. Finally we compare three centralities by means of a real network and real data on the city of Rome (Italy).

A novel graph matrix representation: sequence of neighbourhood matrices with an application

In the study of network optimization, finding the shortest path minimizing time/distance/cost from a source node to a destination node is one of the fundamental problems. Our focus here is to find the shortest path between any pair of nodes in a given undirected unweighted simple graph with the help of the sequence of powers of neighbourhood matrices. The authors recently introduced the concept of neighbourhood matrix as a novel representation of graphs using the neighbourhood sets of the vertices. In this article, an extension of the above work is presented by introducing a sequence of matrices, referred to as the sequence of powers of $${\mathcal {NM}}(G)$$ . It is denoted it by $${\mathcal {NM}}^{\{l\}}(G) = [\eta _{ij}^{\{l\}}], 1\le l \le k(G)$$ , where k(G) is called the iteration number, $$k(G)=\lceil \log _{2} {\hbox {diameter}}(G)\rceil$$ . As this sequence of matrices captures the distance between the nodes profoundly, we further develop the technique and present several characterizations. Based on the theoretical results, we present an algorithm to find the shortest path between any pair of nodes in a given graph. The proposed algorithm and the claims therein are formally validated through simulations on synthetic data and the real network data from Facebook. The empirical results are quite promising with our algorithm having best running time among all the existing well-known shortest path algorithms for the considered graph classes.

A Modified Multi-objective Particle Swarm Optimizer-Based Lévy Flight: An Approach Toward Intrusion Detection in Internet of Things

The emerging of the Internet of things (IoT), and more, the advent of the Internet of everything have revolutionized the computer networks industry. The high diversity of IoT devices, its protocols and standards, and its limited computational resources have led to the appearance of novel security challenges. Hence, the traditional security countermeasures of encryption and authentication are insufficient. Promoting the network security is a fundamental concern for practitioners for safeguarding their economical and industrial strategies. Intrusion detection systems (IDSs) are the major solutions for protecting Internet-connected frameworks at the network-level. But, more importantly, is how to convert the traditional IDSs into intelligent IDSs that resemble the intelligent IoT. This paper presents a new approach for converting the traditional IDSs into smart, evolutionary, and multi-objective IDSs for IoT networks. Moreover, this article presents a modified algorithm for IDSs that tackles the problem of feature selection. The modified algorithm stands on the integration of multi-objective particle swarm optimization with Levy flight randomization component (MOPSO-Levy); the modified MOPSO-Levy has been tested on real IoT network data that is drawn from UCI repository. MOPSO-Levy has achieved superior performance results when compared with state-of-the-art evolutionary multi-objective algorithms.

A Method for Finding Constrained Driver Nodes in Target Control of Network

One of the ultimate goals of studying network dynamics and its properties is to control it. In the past 20 years, with the joint efforts of many scientists, the theory of controlling a whole network through a small set of nodes, which are called driver nodes (DN) has made great progress. However, the full control of networks may be neither feasible nor necessary in some real systems with huge size and high complexity, which motivate scientists to explore target control theory, i.e., the efficient control for a subset of nodes in a network through a small node set. And in a real network, there is another common situation, which is not each node can be easily accessed. Therefore, it is meaningful to explore the target control strategy under the condition that the driver nodes are constrained. In this paper, an effective method is proposed to make more DN be included in the constrained node set (CNs). We adopt the strategy of greedy algorithm to gradually constrain DN into CNs in the iterative process of target control. In each iteration, we will adjust the strategy of how to choose driver nodes according to some network properties. And a few experiments were presented to prove that the proposed method can make DN into CNs effectively in both Scale-free networks (SF) and Erdös-Rényi (ER) networks. Then, we explored the performance of this method in the case of local and random selection of target nodes, respectively. In addition, some factors that will affect the effects of this method were also explored in this paper. In the end, this method is proved valid through the verification of the real network data sets.

Deep Learning-Based DDoS-Attack Detection for Cyber–Physical System Over 5G Network

With the advent of 5G, cyber-physical systems (CPSs) employed in the vertical industries and critical infrastructures will depend on the cellular network more than ever; making their attack surface wider. Hence, guarding the network against cyberattacks is critical not only for its primary subscribers but to prevent it from being exploited as a proxy to attack CPSs. In this article, we propose a consolidated framework, by utilizing deep convolutional neural networks (CNNs) and real network data, to provide early detection for distributed denial-of-service (DDoS) attacks orchestrated by a botnet that controls malicious devices. These puppet devices individually perform silent call, signaling, SMS spamming, or a blend of these attacks targeting call, Internet, SMS, or a blend of these services, respectively, to cause a collective DDoS attack in a cell that can disrupt CPSs' operations. Our results demonstrate that our framework can achieve higher than $91\%$ normal and underattack cell detection accuracy.

Hydra: a method for strain-minimizing hyperbolic embedding of network- and distance-based data

Abstract We introduce hydra (hyperbolic distance recovery and approximation), a new method for embedding network- or distance-based data into hyperbolic space. We show mathematically that hydra satisfies a certain optimality guarantee: it minimizes the ‘hyperbolic strain’ between original and embedded data points. Moreover, it is able to recover points exactly, when they are contained in a low-dimensional hyperbolic subspace of the feature space. Testing on real network data we show that the embedding quality of hydra is competitive with existing hyperbolic embedding methods, but achieved at substantially shorter computation time. An extended method, termed hydra+, typically outperforms existing methods in both computation time and embedding quality.

Bias reduction in the population size estimation of large data sets

Estimation of the population size of large data sets and hard to reach populations can be a significant problem. For example, in the military, manpower is limited and the manual processing of large data sets can be time consuming. In addition, accessing the full population of data may be restricted by factors such as cost, time, and safety. Four new population size estimators are proposed, as extensions of existing methods, and their performances are compared in terms of bias with two existing methods in the big data literature. These would be particularly beneficial in the context of time-critical decisions or actions. The comparison is based on a simulation study and the application to five real network data sets (Twitter, LiveJournal, Pokec, Youtube, Wikipedia Talk). Whilst no single estimator (out of the four proposed) generates the most accurate estimates overall, the proposed estimators are shown to produce more accurate population size estimates for small sample sizes, but in some cases show more variability than existing estimators in the literature.

Extracting Travel Demand for Emergency Situations Using Location-Based Social Network Data

This paper presents the concept of a modular personalized recommendation system, based on existing social networks. The system offers the users in the affected area to evacuate to their nearest best friend or family member who is outside the danger zone. The output of the system is a near real-time OD matrix, which constitutes an input for transportation evacuation models. A preliminary test case shows success in implementing the system on anonymous real location-based social network data.