Complex Network Model Research Articles

Network topologies for maximal organismal health span and lifespan.

The population dynamics of human health and mortality can be jointly captured by complex network models using scale-free network topology. To validate and understand the choice of scale-free networks, we investigate which network topologies maximize either lifespan or health span. Using the Generic Network Model (GNM) of organismal aging, we find that both health span and lifespan are maximized with a "star" motif. Furthermore, these optimized topologies exhibit maximal lifespans that are not far above the maximal observed human lifespan. To approximate the complexity requirements of the underlying physiological function, we then constrain network entropies. Using non-parametric stochastic optimization of network structure, we find that disassortative scale-free networks exhibit the best of both lifespan and health span. Parametric optimization of scale-free networks behaves similarly. We further find that higher maximum connectivity and lower minimum connectivity networks enhance both maximal lifespans and health spans by allowing for more disassortative networks. Our results validate the scale-free network assumption of the GNM and indicate the importance of disassortativity in preserving health and longevity in the face of damage propagation during aging. Our results highlight the advantages provided by disassortative scale-free networks in biological organisms and subsystems.

Vulnerable Line Identification of Cascading Failure in Power Grid Based on New Electrical Betweenness

In this brief, a new electrical betweenness method is proposed to identify the vulnerable lines of the power grid. This method can better balance the accuracy and efficiency of identifying vulnerable lines. The proposed power percentage of generator-load is used as an indicator to evaluate the stability of power grid, and the dynamic influence of generator and load node removal on power grid is considered. The nonlinear cascading failure model of complex networks with overload and weighted edges is introduced, and the three states of normal-failure-overload of lines are fully considered. The model used is more in line with the actual power system. The simulation analysis on IEEE 39 and 118-bus system shows that the change of power loss caused by removing a certain proportion of vulnerable lines and deliberately attacking high new electrical betweenness in turn is compared. The feasibility and effectiveness of this method in identifying vulnerable lines in cascading failure model can be verified.

A Lightweight YOLOv5 Optimization of Coordinate Attention

As Machine Learning technologies evolve, there is a desire to add vision capabilities to all devices within the IoT in order to enable a wider range of artificial intelligence. However, for most mobile devices, their computing power and storage space are affected by factors such as cost and the tight supply of relevant chips, making it impossible to effectively deploy complex network models to small processors with limited resources and to perform efficient real-time detection. In this paper, YOLOv5 is studied to achieve the goal of lightweight devices by reducing the number of original network channels. Then detection accuracy is guaranteed by adding a detection head and CA attention mechanism. The YOLOv5-RC model proposed in this paper is 30% smaller and lighter than YOLOv5s, but still maintains good detection accuracy. YOLOv5-RC network models can achieve a good balance between detection accuracy and detection speed, with potential for its widespread use in industry.

On the Internal Synergistic Mechanism of Operating System of Beijing’s High-Technology Industry Chain: Evidence from Science and Technology Service Industry

Based on the then high-tech industry policy documents issued by Beijing in 2017, this paper builds a complex network model to reflect the internal structure of Beijing’s High-Technology Industry Chain (BHIC), and then analyzes the coevolutionary mechanisms of ten top high-technology industries and their detailed sectors. Especially for the science and technology service (S&T) industry, this paper measures its function and status in the system and studies its concrete approach to stabilize the development of the industry chain. Finally, policy suggestions to promote the development of the S&T service industry in specialization, networking, and scale are put forward.

A Novel Link Prediction Framework Based on Gravitational Field

Currently, most researchers only utilize the network information or node characteristics to calculate the connection probability between unconnected node pairs. Therefore, we attempt to project the problem of connection probability between unconnected pairs into the physical space calculating it. Firstly, the definition of gravitation is introduced in this paper, and the concept of gravitation is used to measure the strength of the relationship between nodes in complex networks. It is generally known that the gravitational value is related to the mass of objects and the distance between objects. In complex networks, the interrelationship between nodes is related to the characteristics, degree, betweenness, and importance of the nodes themselves, as well as the distance between nodes, which is very similar to the gravitational relationship between objects. Therefore, the importance of nodes is used to measure the mass property in the universal gravitational equation and the similarity between nodes is used to measure the distance property in the universal gravitational equation, and then a complex network model is constructed from physical space. Secondly, the direct and indirect gravitational values between nodes are considered, and a novel link prediction framework based on the gravitational field, abbreviated as LPFGF, is proposed, as well as the node similarity framework equation. Then, the framework is extended to various link prediction algorithms such as Common Neighbors (CN), Adamic-Adar (AA), Preferential Attachment (PA), and Local Random Walk (LRW), resulting in the proposed link prediction algorithms LPFGF-CN, LPFGF-AA, LPFGF-PA, LPFGF-LRW, and so on. Finally, four real datasets are used to compare prediction performance, and the results demonstrate that the proposed algorithmic framework can successfully improve the prediction performance of other link prediction algorithms, with a maximum improvement of 15%.

Hitting Times of Random Walks on Edge Corona Product Graphs

Abstract Graph products have been extensively applied to model complex networks with striking properties observed in real-world complex systems. In this paper, we study the hitting times for random walks on a class of graphs generated iteratively by edge corona product. We first derive recursive solutions to the eigenvalues and eigenvectors of the normalized adjacency matrix associated with the graphs. Based on these results, we further obtain interesting quantities about hitting times of random walks, providing iterative formulas for two-node hitting time, as well as closed-form expressions for the Kemeny’s constant defined as a weighted average of hitting times over all node pairs, as well as the arithmetic mean of hitting times of all pairs of nodes.

A Network Science Perspective of Graph Convolutional Networks: A Survey

The mining and exploitation of graph structural information have been the focal points in the study of complex networks. Traditional structural measures in Network Science focus on the analysis and modelling of complex networks from the perspective of network structure, such as the centrality measures, the clustering coefficient, and motifs and graphlets, and they have become basic tools for studying and understanding graphs. In comparison, graph neural networks, especially graph convolutional networks (GCNs), are particularly effective at integrating node features into graph structures via neighbourhood aggregation and message passing, and have been shown to significantly improve the performances in a variety of learning tasks. These two classes of methods are, however, typically treated separately with limited references to each other. In this work, aiming to establish relationships between them, we provide a network science perspective of GCNs. Our novel taxonomy classifies GCNs from three structural information angles, i.e., the layer-wise message aggregation scope, the message content, and the overall learning scope. Moreover, as a prerequisite for reviewing GCNs via a network science perspective, we also summarise traditional structural measures and propose a new taxonomy for them. Finally and most importantly, we draw connections between traditional structural approaches and graph convolutional networks, and discuss potential directions for future research.

Maximizing portfolio profitability during a cryptocurrency downtrend: A Bitcoin Blockchain transaction-based approach

The volatile and unpredictable nature of the cryptocurrency market makes it particularly challenging to make profitable investment decisions. different machine learning-based techniques have been employed for forecasting cryptocurrency value. However, although some works have addressed incorporating the Blockchain transactions’ data into the analysis, none of them has provided a hybrid solution, including features obtained through complex network modeling. In this paper, we investigated the use of machine learning and complex network techniques to improve the profitability of a cryptocurrency portfolio during a downtrend period. We extracted features through a complex network-building methodology based on the Bitcoin blockchain transactions, merged them with the historical cryptocurrency values, and generated the predictions using different machine-learning models. The results indicated that incorporating complex network features improved the performance in retaining the initial capital at the end of the experiment, leading to an increment of 7.09% and 4.33% for the CNN and LSTM models, respectively. Our findings suggest that the proposed method enhanced the performance of cryptocurrency investment strategies during downtrend periods.

Spatiotemporal distribution prediction of coughing airflow at mouth based on machine learning - Part I: Study on boundary conditions at mouth in numerical simulation of cough airflow

In the post epidemic era, the movement and distribution of pathogenic airflow and droplets produced by cough in the building space have been widely studied. Due to the limitations of research methods, there are few detailed research data on the temporal and spatial distribution of boundary conditions during cough, which is the basis of research and the key boundary conditions of computer simulation. Previous experiments have obtained cough airflow velocity distribution away from the mouth. This study aims to infer detailed data at mouth for CFD boundary conditions based on these experimental data. This is the first part of the research. Based on experiments, the types of parameters contained in the boundary conditions near the mouth during coughing are discussed. The main parameters are determined, including the maximum velocity of the mouth air flow, and the distribution function of the ejected air flow, among others, and the approximate value range. Different parameter combinations are used as boundary conditions for simulation, and with various combinations, database of conditions are obtained. Preliminary machine learning is performed on these databases, and boundary condition data consistent with experimental results are inferred. The study demonstrates that when the velocity distribution of the air flow at mouth satisfies the normal distribution function on the central vertical two-dimensional profile, the maximum velocity of the mouth air flow is 15m/s. Part 2 will use the complex neural network model to fit and infer more accurate boundary condition. The findings of this study can provide more accurate boundary conditions for simulating pathogenic airflow, as well as a supplementary database for epidemiological research.

Transmission mechanism and effect of international nonferrous metal price fluctuation to China’s industrial chain:Based on a two-layer complex network model

Transmission mechanism and effect of international nonferrous metal price fluctuation to China’s industrial chain:Based on a two-layer complex network model

Comparing the centrality symptoms of major depressive disorder samples across junior high school students, senior high school students, college students and elderly adults during city lockdown of COVID-19 pandemic—A network analysis

IntroductionRecently, in the view of network analysis, depression has been conceptualized as a complex and dynamic network model combining individual symptoms. To date, no studies have systematically examined and compared depressive symptom networks across different populations. MethodsA total of 36,105 participants were recruited and asked to complete the Patient Health Questionnaire-9 among junior high school students, senior high school students, college students, and elderly adults who were more susceptible to depression during the COVID-19 lockdown in China. In the analysis, we applied the optimal cutoff score ≥ 8 for students and a score ≥ 6 for elderly adults to identify 5830 participants who were likely to be depressed. The index of “strength” was used to identify central symptoms in the network structure. ResultsThe results showed that Sad Mood was the most central symptom among junior high school students, senior high school students, and college students, but the most central symptom in the elderly was Guilt. Among the top three central symptoms, Suicide Ideation was unique to senior high school students, while Anhedonia was most prevalent among college students. Guilt - Suicide Ideation, Anhedonia – Energy, Anhedonia - Sad Mood, and Sleep – Energy showed the strongest association among junior and senior high school students, college students, and elderly adults, respectively. NCT (i.e., Network Comparison Test) suggested that the network's global connectivity was ultimately inconsistent, but the network structure remained roughly intact. ConclusionIn treatment, targeting central symptoms may be critical to alleviating depression.

Robustness of Cloud Manufacturing System Based on Complex Network and Multi-Agent Simulation

Cloud manufacturing systems (CMSs) are networked, distributed and loosely coupled, so they face great uncertainty and risk. This paper combines the complex network model with multi-agent simulation in a novel approach to the robustness analysis of CMSs. Different evaluation metrics are chosen for the two models, and three different robustness attack strategies are proposed. To verify the effectiveness of the proposed method, a case study is then conducted on a cloud manufacturing project of a new energy vehicle. The results show that both the structural and process-based robustness of the system are lowest under the betweenness-based failure mode, indicating that resource nodes with large betweenness are most important to the robustness of the project. Therefore, the cloud manufacturing platform should focus on monitoring and managing these resources so that they can provide stable services. Under the individual server failure mode, system robustness varies greatly depending on the failure behavior of the service provider: Among the five service providers (S1-S5) given in the experimental group, the failure of Server 1 leads to a sharp decline in robustness, while the failure of Server 2 has little impact. This indicates that the CMS can protect its robustness by identifying key servers and strengthening its supervision of them to prevent them from exiting the platform.

Individual-Based Modelling of Animal Brucellosis Spread with the Use of Complex Networks

Article Individual-Based Modelling of Animal Brucellosis Spread with the Use of Complex Networks E. R. Pinto 1, E. G. Nepomuceno 2,*, and A. S. L. O. Campanharo 1 1 Department of Biostatistics, Institute of Biosciences São Paulo State University, Botucatu, São Paulo 18618-970, Brazil 2 Department of Electronic Engineering and Centre for Ocean Energy Research, Maynooth University, Maynooth, Ireland * Correspondence: erivelton.nepomuceno@mu.ie Received: 12 October 2022 Accepted: 23 November 2022 Published: 22 December 2022 Abstract: The principal purpose of this work was to study the spread of brucellosis in the state of São Paulo with the help of the complex network theory and to propose control measures for its eradication. For this, the scale-free model of complex networks, widely known in the literature, was used. The effect of vaccination was verified in each of the municipalities in the state of São Paulo and it was observed that when heterogeneity is not taken into account, vaccination becomes ineffective for the eradication of the disease.

Class of models for random hypergraphs.

Despite the recently exhibited importance of higher-order interactions for various processes, few flexible (null) models are available. In particular, most studies on hypergraphs focus on a small set of theoretical models. Here, we introduce a class of models for random hypergraphs which displays a similar level of flexibility of complex network models and where the main ingredient is the probability that a node belongs to a hyperedge. When this probability is a constant, we obtain a random hypergraph in the same spirit as the Erdos-Renyi graph. This framework also allows us to introduce different ingredients such as the preferential attachment for hypergraphs, or spatial random hypergraphs. In particular, we show that for the Erdos-Renyi case there is a transition threshold scaling as 1/sqrt[EN] where N is the number of nodes and E the number of hyperedges. We also discuss a random geometric hypergraph which displays a percolation transition for a threshold distance scaling as r_{c}^{*}∼1/sqrt[E]. For these various models, we provide results for the most interesting measures, and also introduce new ones in the spatial case for characterizing the geometrical properties of hyperedges. These different models might serve as benchmarks useful for analyzing empirical data.

Functional importance evaluation approach for cloud manufacturing services based on complex network and evidential reasoning rule

In the uncertain cloud environment, manufacturing services monitoring is effective to ensure the normal function of cloud manufacturing service system (CMSS). However, continuously monitoring all of the manufacturing services is impractical since it is resource consuming. One feasible method is to prioritize the allocation of monitoring resources to the important services. Therefore, we propose an approach for evaluating the functional importance of manufacturing services based on complex network and evidential reasoning (ER) rule. Firstly, a domain-oriented cloud manufacturing service complex network (DoCMSCN) model is constructed and elaborated. Secondly, based on the idea of multi-granularity and multi-indicator, an evaluation model for the importance of manufacturing services is presented. Different centrality indicators of the DoCMSCN in different functional granularities are obtained and transformed into evaluation evidence. Then the ER algorithm is applied to fused the evaluation results. In the fusion process, the reliability and weight of each piece of evidence are fully considered to improve the fusion accuracy. The experimental results of vertical elevator design services show the proposed approach can effectively find the important manufacturing services and superior than the existing ones. It can facilitate the decision-making of monitoring strategy in conditions of limited resources from the functional perspective. Finally, we develop a prototype monitoring system for vertical elevator design services.

Incorporating circuit theory, complex networks, and carbon offsets into the multi-objective optimization of ecological networks: A case study on karst regions in China

Ecological networks (ENs) are important for maintaining regional ecological security and improving ecosystem service capacity. Therefore, it is important to consider spatial topological relationships and the need for carbon neutrality in the optimization of ENs. The optimization of EN structures and functions should be a systematic multi-objective process. This study combined landscape ecology, the complex network model, and spatial analysis technology, proposing an ecological barrier–topological feature–carbon offset-based (BTC-based) model research framework to deeply analyze and optimize ENs. First, the EN was preliminarily identified using the circuit theory model and the ecological barrier areas were simulated. Then, based on the complex network model, an undirected and unweighted complex network was established and its topological structural characteristics were analyzed. By constructing the estimation model for the carbon offset rate, its spatial distribution characteristics were calculated and interpreted. Finally, the BTC-based model was used to optimize the EN. Based on this model, Guizhou Province (a typical karst region in China) was taken as the research area for this study. The results showed that ecological improvement and degradation coexisted, and measures must be taken to optimize the EN. Based on the 2018 EN, the areas that required optimization were identified. Additionally, a restoration strategy was proposed for edge addition optimization for ecological corridors to promote the ecological protection and quality improvement of ecological sources and corridors. This approach would ensure the accurate implementation of ecological restoration projects and the effective management of ecosystems. The model proposed in this study also provides methodological support for EN optimization in other types of ecological regions.

Critical properties of the Anderson transition on random graphs: Two-parameter scaling theory, Kosterlitz-Thouless type flow, and many-body localization

The Anderson transition in random graphs has raised great interest, partly because of its analogy with the many-body localization (MBL) transition. Unlike the latter, many results for random graphs are now well established, in particular the existence and precise value of a critical disorder separating a localized from an ergodic delocalized phase. However, the renormalization group flow and the nature of the transition are not well understood. In turn, recent works on the MBL transition have made the remarkable prediction that the flow is of Kosterlitz-Thouless type. Here we show that the Anderson transition on graphs displays the same type of flow. Our work attests to the importance of rare branches along which wave functions have a much larger localization length $\xi_\parallel$ than the one in the transverse direction, $\xi_\perp$. Importantly, these two lengths have different critical behaviors: $\xi_\parallel$ diverges with a critical exponent $\nu_\parallel=1$, while $\xi_\perp$ reaches a finite universal value ${\xi_\perp^c}$ at the transition point $W_c$. Indeed, $\xi_\perp^{-1} \approx {\xi_\perp^c}^{-1} + \xi^{-1}$, with $\xi \sim (W-W_c)^{-\nu_\perp}$ associated with a new critical exponent $\nu_\perp = 1/2$, where $\exp( \xi)$ controls finite-size effects. The delocalized phase inherits the strongly non-ergodic properties of the critical regime at short scales, but is ergodic at large scales, with a unique critical exponent $\nu=1/2$. This shows a very strong analogy with the MBL transition: the behavior of $\xi_\perp$ is identical to that recently predicted for the typical localization length of MBL in a phenomenological renormalization group flow. We demonstrate these important properties for a smallworld complex network model and show the universality of our results by considering different network parameters and different key observables of Anderson localization.

The New Generation Brain-Inspired Sparse Learning: A Comprehensive Survey

In recent years, the enormous demand for computing resources resulting from massive data and complex network models has become the limitation of deep learning. In the large-scale problems with massive samples and ultrahigh feature dimensions, sparsity has gradually drawn much attention from academia and the industrial field. In this article, the new generation of brain-inspired sparse learning is reviewed comprehensively. First, sparse cognition learning is introduced from the visual biology mechanism to modeling for the natural image. Second, the sparse representation algorithms are summarized to sort out the research progress of sparse learning. Third, the relevant research on sparse feature selection learning is reviewed. Then, the sparse deep networks and applications are summed up. Last but not least, ten public issues and challenges of sparse learning are discussed. By investigating the development process of sparse learning, this article summarizes the advantages, disadvantages, limitations, and future research directions of the algorithm, which can help readers conduct further study.

Spectral induced polarization of heterogeneous non-consolidated clays

SUMMARY Clays are ubiquitously located in the Earth’s near surface and have a high impact on the subsurface permeability. Most geo-electrical characterizations of clays do not take into account the heterogeneous nature of clay geological media. We want to better understand the influence of heterogeneities on the geo-electrical signature, thus we collected a data set of spectral induced polarization (SIP) of artificial heterogeneous non-consolidated clay samples. The samples are made of illite and red montmorillonite in a parallel and perpendicular disposition (with respect to the applied electric field). Another sample is a homogeneous mixture composed of the same volumetric fraction of illite and red montmorillonite. For all the samples, the polarization is dominated by the red montmorillonite, given by the shape of the spectra (presence or lack of a peak at a particular frequency). We compared the experimental data with classical mixing laws and complex conductance network models to test how to better predict the SIP signature of such mixtures when the SIP spectra of the two components are known. The real conductivity is better predicted by the mixing laws, but the shape of the spectra (presence of polarization peaks at particular frequencies) is best predicted by the conductance network models. This study is a step forward towards a better characterization of heterogeneous clay systems using SIP.

Identification of City Hotspots by Analyzing Telecom Call Detail Records Using Complex Network Modeling

Many interesting observations are made from the analysis of Telecom bigdata called call detail records (CDR) on various aspects of city dynamics. CDR is a low-cost, real-time, and clean (no human error) data with abundant availability to the state administrations. However, due to its massive size, the processing of raw CDR is not feasible most of the time. Graph-theoretic approaches to network analysis are enriched with efficient algorithms and heuristics for modeling and analyzing various generic problems. Once a CDR dataset is modeled as a communication network with aggregation of raw data spatially and temporally, network analysis tools yield significant results. A heightened communication activity (call/SMS/Internet) originating from a location/area to many other areas signify the importance of that originating area. In this paper, we have modeled the telecommunication data of two Italian cities (Milan and Trento) as a weighted network. The modeled network from raw CDR becomes very dense due to the large number of edges. We have pre-processed and thresholded the available CDR bigdata to generate many smaller-sized sparse networks by removing less significant edges. Different well-known node-centrality metrics (weighted degree, weighted k-shell) and our proposed modified edge weight degree neighborhood (EwDN) method are applied on the modeled network to identify the top-k most important nodes representing the hotspot locations in the city. We also use five days data and compare changes in the hotspot patterns. Finally, we validate our findings (hotspot locations) with ground truth from other sources using projected geo-location tagging on QGIS (An open source Geographic Information System application tool). To overcome the availability of limited ground truth, we use the standard Susceptible–Infectious–Recovered (SIR) model as a benchmark and compare various methods using Kendall’s rank correlation. Results indicate that the proposed network modeling, size reduction by edge-weight threshold, and application of complex network approaches including the proposed modified EwDN method identify important places in almost real-time with much less computational overhead.