Optimizing Multilayer Networks Through Time-Dependent Decision-Making: A Comparative Study.

BackgroundThe HIV-1 molecular network is an innovative tool, using gene sequences to understand transmission attributes and complementing social and sexual network studies. While previous research focused on static network characteristics, recent studies’ emphasis on dynamic features enhances our understanding of real-time changes, offering insights for targeted interventions and efficient allocation of public health resources.ObjectiveThis study aims to identify the dynamic changes occurring in HIV-1 molecular transmission networks and analyze the primary influencing factors driving the dynamics of HIV-1 molecular networks.MethodsWe analyzed and compared the dynamic changes in the molecular network over a specific time period between the baseline and observed end point. The primary factors influencing the dynamic changes in the HIV-1 molecular network were identified through univariate analysis and multivariate analysis.ResultsA total of 955 HIV-1 polymerase fragments were successfully amplified from 1013 specimens; CRF01_AE and CRF07_BC were the predominant subtypes, accounting for 40.8% (n=390) and 33.6% (n=321) of the specimens, respectively. Through the analysis and comparison of the basic and terminal molecular networks, it was discovered that 144 sequences constituted static molecular networks, and 487 sequences contributed to the formation of dynamic molecular networks. The findings of the multivariate analysis indicated that the factors occupation as a student, floating population, Han ethnicity, engagement in occasional or multiple sexual partnerships, participation in anal sex, and being single were independent risk factors for the dynamic changes observed in the HIV-1 molecular network, and the odds ratio (OR; 95% CIs) values were 2.63 (1.54-4.47), 1.83 (1.17-2.84), 2.91 (1.09-7.79), 1.75 (1.06-2.90), 4.12 (2.48-6.87), 5.58 (2.43-12.80), and 2.10 (1.25-3.54), respectively. Heterosexuality and homosexuality seem to exhibit protective effects when compared to bisexuality, with OR values of 0.12 (95% CI 0.05-0.32) and 0.26 (95% CI 0.11-0.64), respectively. Additionally, the National Eight-Item score and sex education experience were also identified as protective factors against dynamic changes in the HIV-1 molecular network, with OR values of 0.12 (95% CI 0.05-0.32) and 0.26 (95% CI 0.11-0.64), respectively.ConclusionsThe HIV-1 molecular network analysis showed 144 sequences in static networks and 487 in dynamic networks. Multivariate analysis revealed that occupation as a student, floating population, Han ethnicity, and risky sexual behavior were independent risk factors for dynamic changes, while heterosexuality and homosexuality were protective compared to bisexuality. A higher National Eight-Item score and sex education experience were also protective factors. The identification of HIV dynamic molecular networks has provided valuable insights into the characteristics of individuals undergoing dynamic alterations. These findings contribute to a better understanding of HIV-1 transmission dynamics and could inform targeted prevention strategies.

Link prediction in multilayer networks using weighted reliable local random walk algorithm

Econometricians are increasingly working with high-dimensional networks and their dynamics. Econometricians, however, are often confronted with unforeseen changes in network dynamics. In this article, we develop a method and the corresponding algorithm for monitoring changes in dynamic networks. We characterize two types of changes, edge-initiated and node-initiated, to feature the complexity of networks. The proposed approach accounts for three potential challenges in the analysis of networks. First, networks are high-dimensional objects causing the standard statistical tools to suffer from the curse of dimensionality. Second, any potential changes in social networks are likely driven by a few nodes or edges in the network. Third, in many dynamic network applications such as monitoring network connectedness or its centrality, it will be more practically applicable to detect the change in an online fashion than the offline version. The proposed detection method at each time point projects the entire network onto a low-dimensional vector by taking the sparsity into account, then sequentially detects the change by comparing consecutive estimates of the optimal projection direction. As long as the change is sizeable and persistent, the projected vectors will converge to the optimal one, leading to a jump in the sine angle distance between them. A change is therefore declared. Strong theoretical guarantees on both the false alarm rate and detection delays are derived in a sub-Gaussian setting, even under spatial and temporal dependence in the data stream. Numerical studies and an application to the social media messages network support the effectiveness of our method.

BackgroundMild Cognitive Impairment (MCI) is an intermediate stage between the expected cognitive decline of normal aging and Alzheimer’s disease (AD). Its management is crucial for it helps intervene and slow the progression of cognitive decline to AD. However, the understanding of the MCI mechanism is not completely clear. As working memory (WM) damage is a common symptom of MCI, this study focused on the core stage of WM, i.e., the memory retrieval stage, to investigate information processing and the causality relationships among brain regions based on electroencephalogram (EEG) signals.Method21 MCI and 20 normal cognitive control (NC) participants were recruited. The delayed matching sample paradigm with two different loads was employed to evaluate their WM functions. A time-varying network based on the Adaptive transfer function (ADTF) was constructed on the EEG of the memory retrieval trials.to perform the dynamic brain network analysis.ResultsOur results showed that: (a) Behavioral data analysis: there were significant differences in accuracy and accuracy / reaction time between MCI and NC in tasks with memory load capacity of low load-four and high load-six, especially in tasks with memory load capacity of four. (b) Dynamic brain network analysis: there were significant differences in the dynamic changes of brain network patterns between the two groups during the memory retrieval stage of the WM task. Specifically, in low load WM tasks, the dynamic brain network changes of NC were more regular to accommodate for efficient information processing, with important core nodes showing a transition from bottom to up, while MCI did not display a regular dynamic brain network pattern. Further, the brain functional areas associated with low load WM disorders were mainly located in the left prefrontal lobe (FC1) and right occipital lobe (PO8). Compared with low load WM task, during the high load WM task, the dynamic brain network changes of NC during the memory retrieval stage were regular, and the core nodes exhibited a consistent transition phenomenon from up to bottom to up, which were not observed in MCI.ConclusionsBehavioral data in the low load WM task paradigm and abnormal electrophysiological signals in the left prefrontal (FC1) and right occipital lobes (PO8) could be used for MCI diagnosis. This is the first time based on large-scale dynamic network methods to investigate the dynamic network patterns of MCI memory retrieval stages under different load WM tasks, providing a new perspective on the neural mechanisms of WM deficits in MCI patients and providing some reference for the clinical intervention treatment of MCI-WM memory disorders.

A complex network is an explicit model for a real-world system such as technological networks, social networks, business networks, and biological networks. The social network is an Internet-based media for socially relevant activities like stay connected with families and friends, colleagues, and customers, for socializing, business, or both. The key nodes, usually called central nodes, capable of measuring the performance of various social network applications. Identifying influencing nodes is primary research for any network analysis research. Degree centrality, a locally computed metric, is simple to compute but not persuasive. The global metrics like betweenness centrality and PageRank are only useful for the systems with a simple structure, but incur a high computational cost with the addition of layers. This paper proposes a novel metric m-PageRank for ranking nodes in a multi-layer complex network. The m-PageRank is an advancement of PageRank. It integrates the existence of the rank of each layer from where the connection connects. The proposed metric was validated through simulations performed over various multilayer networks. The result shows that m-PageRank computes the rank of each node accurately. We observe that the comparison with state-of-the-art metrics demonstrated the suitability of the proposed metric.

Time-lapse visualization of spatial and temporal patterns of stream network dynamics.

Many complex, real-world dynamic systems need to be described using multilayer networks. Reconstruction of the multilayer nonlinear dynamic network structure from the observed streaming data is one of the core issues of dynamic system research. Although many methods have been proposed to reconstruct multilayer nonlinear dynamic networks from static data, they cannot reconstruct the multilayer network from streaming data. Because of the complexity of the problem, there is no nonlinear dynamic network reconstruction model and its optimization algorithm suitable for streaming data. To resolve this limitation, in this paper, a streaming data-driven multilayer dynamic network reconstruction framework called online multilayer network reconstruction (OMNR) is proposed. OMNR first establishes an online multilayer nonlinear dynamics network reconstruction model for streaming data and then proposes an online learning method based on follow-the-regularized-leader to optimize the reconstruction model. It considers only one case at a time to update the network structure; this is suitable for processing streaming data. Moreover, the space complexity of OMNR is very low and suitable for big-data tasks. Finally, using the Lorenz and Rossler dynamic systems of the multilayer network, it is verified that OMNR can solve the problem of multilayer dynamic network reconstruction driven by streaming data and fill the gap in the technology of multilayer nonlinear dynamic network reconstruction driven by streaming data.

The real-world complex networks, such as biological, transportation, biomedical, web, and social networks, are usually dynamic and change over time. The communities which reflect the substructures hidden in the networks usually overlap each other, and detecting overlapping communities in the dynamic complex networks is a challenging task. Prior researchers have applied multiobjective optimization method to the detection of dynamic overlapping communities and achieved some excellent results. However, in terms of multiobjective processing, the prior studies all adopt the decomposition method based on weight parameters, and different weight parameters or different parameter values can easily affect the community detection results which further results in the uneven distribution of the detected results in the target space. To solve the above problems, a hybrid algorithm, that is, Collaborative Particle Swarm multiobjective Optimization-based Dynamic Overlapping Community Detection (CPSO-DOCD) algorithm is proposed in this paper. First, to improve the diversity of particles, the encoding/decoding of the particle and the cross inheritance and the variation of particle are redefined first based on label propagation. In each network snapshot, multiple particle swarms are initialized based on Community Overlap Propagation Algorithm (COPRA) to generate particles with uniform distribution. Multiple different objective functions are optimized using multiple particle swarms respectively to avoid the incorrect selection of weight parameters. In addition, a reference-point-based is adopted in the particle selecting stage to solve the uneven distribution of detected results in the target space. Second, a node label entropy-based particle swarm algorithm is proposed to improve the accuracy of community detection of current network snapshots. Finally, when one snapshot switches to another over time, a migration strategy based on COPRA local-search and clique generation is utilized to adjust the prior community detection results, which enables the former results can be adapted to the new network snapshots. The experiments are implemented based on four dynamic networks which are Cit-HepPh, Cit-HepTh, Emailed-EU-core-temporal, and CollegeMsg. The hypervolume value of the overlapping community detection result obtained by CPSO-DOCD is 0.5%–2% higher than MDOA, MCMOEA, SLPAD, and iLCD. Furthermore, CPSO-DOCD also performed better than MDOA, MCMOEA, SLPAD, and iLCD on C-metric values, and CPSO-DOCD can approach approximately to the Pareto frontier.

Multilayer networks are formed by several networks that interact with each other and co-evolve. Multilayer networks include social networks, financial markets, transportation systems, infrastructures and molecular networks and the brain. The multilayer structure of these networks strongly affects the properties of dynamical and stochastic processes defined on them, which can display unexpected characteristics. For example, interdependencies between different networks of a multilayer structure can cause cascades of failure events that can dramatically increase the fragility of these systems; spreading of diseases, opinions and ideas might take advantage of multilayer network topology and spread even when its single layers cannot sustain an epidemic when taken in isolation; diffusion on multilayer transportation networks can significantly speed up with respect to diffusion on single layers; finally, the interplay between multiplexity and controllability of multilayer networks is a problem with major consequences in financial, transportation, molecular biology and brain networks. This field is one of the most prosperous recent developments of Network Science and Data Science. Multilayer networks include multiplex networks, multi-slice temporal networks, networks of networks, interdependent networks. Multilayer networks are characterized by having a highly correlated multilayer network structure, providing a significant advantage for extracting information from them using multilayer network measures and centralities and community detection methods. The multilayer network dynamics (including percolation, epidemic spreading, diffusion, synchronization, game theory and control) is strongly affected by the multilayer network topology. This book will present a comprehensive account of this emerging field.

We tested whether changes in functional networks predict cognitive decline and conversion from the presymptomatic prodrome to symptomatic disease in familial frontotemporal dementia (FTD). For hypothesis generation, 36 participants with behavioral variant FTD (bvFTD) and 34 controls were recruited from one site. For hypothesis testing, we studied 198 symptomatic FTD mutation carriers, 341 presymptomatic mutation carriers, and 329 family members without mutations. We compared functional network dynamics between groups, with clinical severity and with longitudinal clinical progression. We identified a characteristic pattern of dynamic network changes in FTD, which correlated with neuropsychological impairment. Among presymptomatic mutation carriers, this pattern of network dynamics was found to a greater extent in those who subsequently converted to the symptomatic phase. Baseline network dynamic changes predicted future cognitive decline in symptomatic participants and older presymptomatic participants. Dynamic network abnormalities in FTD predict cognitive decline and symptomatic conversion. We investigated brain network predictors of dementia symptom onset Frontotemporal dementia results in characteristic dynamic network patterns Alterations in network dynamics are associated with neuropsychological impairment Network dynamic changes predict symptomatic conversion in presymptomatic carriers Network dynamic changes are associated with longitudinal cognitive decline.

With the increase in popularity of the social networks, it has become a perennial source of data analytics to mining the abundant information in real-world networks. As the social network is heterogeneous, some of its entities like nodes and edges may be more influential than other entities. It is observed that identification of the most influential entities in large-scale networks has many real-time applications like social network analysis, fraud detection, community detection, traffic control of the transportation network, software-defined network, and many more. Several centrality measures exist to identify the importance of a node in the network. However, betweenness centrality is found to be the most promising one, to investigate a network and the importance of nodes in the network. In the era of big data, the size of the social network is increasing exponentially. Although a number of algorithms exist for identifying betweenness centrality for large-scale networks, very few algorithms attempt to identify the influential nodes in a dynamic network. A single insertion or deletion of a node or edge may drastically change the structure of the network, which limits the performance of algorithms in terms of computational efficiency. In order to accommodate this problem, in this paper, a MapReduce-based incremental parallel algorithm for exploring the influential nodes based on betweenness centrality is proposed. The proposed algorithm is compared with a few other algorithms that are used to compute betweenness centrality in a dynamic network. The major advantage of the proposed algorithm is that it allows the network to be updated by the insertion of an edge. The effectiveness of the proposed algorithm has been critically examined in a distributed environment in terms of execution time by using both real-time and synthetic networks. The experimental results show a significant speedup over the other algorithms.

In recent years, the exponential growth of online social networks as complex networks has presented challenges in expanding networks and forging new connections. Link prediction emerges as a crucial technique to anticipate future relationships among users, leveraging the current network state to address this challenge effectively. While link prediction models on monoplex networks have a well-established history, the exploration of similar tasks on multilayer networks has garnered considerable attention. Extracting topological and multimodal features for weighting links can improve link prediction in weighted complex networks. Meanwhile, establishing reliable and trustworthy paths between users is a useful way to create metrics that convert unweighted to weighted similarity. The local random walk is a widely used technique for predicting links in weighted monoplex networks. The aim of this paper is to develop a semi-local random walk over reliable paths to improve link prediction on a multilayer social network as a complex network, which is denoted as Reliable Multiplex semi-Local Random Walk (RMLRW). RMLRW leverages the semi-local random walk technique over reliable paths, integrating intra-layer and inter-layer information from multiplex features to conduct a trustworthy biased random walk for predicting new links within a target layer of multilayer networks. In order to make RMLRW scalable, we develop a semi-local random walk-based network embedding to represent the network in a lower-dimensional space while preserving its original characteristics. Extensive experimental studies on several real-world multilayer networks demonstrate the performance assurance of RMLRW compared to equivalent methods. Specifically, RMLRW improves the average f-measure of the link prediction by 3.2% and 2.5% compared to SEM-Path and MLRW, respectively.

Event logs can be used to keep track of fine-grained changes in dynamic networks. An event log is a sequence of change events, where each event associates an action (add or remove) with an element (node or edge) and a timestamp. The main benefit of an event log is the ability to construct a static graph (snapshot) of the network at any time point within the lifetime of the dynamic network. We present algorithms for construction of event logs from temporal edge lists for several scenarios. We also present algorithms for construction of static graphs from event logs, eventually using a list of snapshots. We conduct experiments to demonstrate the utility of event logs, and to study the efficiency of static graph construction with respect to the number of used snapshots.

Dynamic Resting-State Network Biomarkers of Antidepressant Treatment Response

This chapter aims to provide discussions of active period selection technique for the cluster-tree type wireless sensor networks (WSNs) employing traffic adaptive IEEE 802.15.4 beacon enabled mode under dynamic network changes, and proposes an autonomous distributed superframe duration (SD) selection scheme to reduce beacon collisions by using simple control mechanism. The dynamic network changes, including network topology changes and frame structure changes, cause severe beacon collisions in the cluster-tree type WSNs and greatly degrade their system performance. To overcome this problem, the proposed scheme autonomously selects an active SD by using beacon reception power monitoring with distributed fashion and also introduces a beacon status notice from sensor nodes (SNs) to their parent cluster heads (CHs) in order to prevent unnecessary SD selection at CHs. To enhance the system performance, this chapter also applies the traffic adaptive distributed backoff mechanism, previously proposed in our recent work, to the autonomous distributed SD selection scheme and investigates an effect of the distributed backoff to the system performance. The results evaluated by computer simulation show that the proposed scheme can improve the transmission performance while keeping the better power consumption performance in cluster-tree type WSNs under cluster mobility environments.