Empirical Networks Research Articles

Social learning mechanisms shape transmission pathways through replicate local social networks of wild birds.

The emergence and spread of novel behaviours via social learning can lead to rapid population-level changes whereby the social connections between individuals shape information flow. However, behaviours can spread via different mechanisms and little is known about how information flow depends on the underlying learning rule individuals employ. Here, comparing four different learning mechanisms, we simulated behavioural spread on replicate empirical social networks of wild great tits and explored the relationship between individual sociality and the order of behavioural acquisition. Our results reveal that, for learning rules dependent on the sum and strength of social connections to informed individuals, social connectivity was related to the order of acquisition, with individuals with increased social connectivity and reduced social clustering adopting new behaviours faster. However, when behavioural adoption depends on the ratio of an individuals' social connections to informed versus uninformed individuals, social connectivity was not related to the order of acquisition. Finally, we show how specific learning mechanisms may limit behavioural spread within networks. These findings have important implications for understanding whether and how behaviours are likely to spread across social systems, the relationship between individuals' sociality and behavioural acquisition, and therefore for the costs and benefits of sociality.

Stock index forecasting based on multivariate empirical mode decomposition and temporal convolutional networks

Modeling and forecasting of stock indexes is an important and challenging work in the field of financial research, which is of great significance for investors to reduce risk on investment and improve investment returns. However, the basic data of the stock index includes five indicators, namely, opening price, highest price, lowest price, closing price and trading volume (COHLV), each of which contains some information related to the future trend. And these indicators are affected by different factors such as politics, economy, psychology and so on, so they have nonlinear, high noise and other complex characteristics. These reasons make the existing methods cannot effectively improve the accuracy of stock index forecasting. In order to solve this problem, we propose a hybrid stock index forecasting model named MEMD-TCN, which is based on multivariate empirical mode decomposition (MEMD) and temporal convolutional networks (TCN). The novelty of the proposed model focuses on decomposition of multivariate time series and using only fully convolutional layers to forecast stock index. The proposed model first uses MEMD to decompose the data of COHLV to obtain subsequences of different fluctuation frequencies of each time series, then inputs subsequences of the same frequency into TCN to predict the subsequences of closing price in the next period. Ultimately, the forecasted values for the closing price are obtained by reconstructing the prediction results of all the subsequences of closing price. To evaluate the performance of the proposed MEMD-TCN model, stock index of several countries that can reflect the overall changes of the market are used. The experimental results verify the notable effectiveness and necessity of the decomposition of multivariate time series. Meanwhile, the results demonstrate the significant superiority of the proposed MEMD-TCN model on accuracy and stability.

Evaluation of Respondent-Driven Sampling Prevalence Estimators Using Real-World Reported Network Degree.

Respondent-driven sampling (RDS) is used to measure trait or disease prevalence in populations that are difficult to reach and often marginalized. The authors evaluated the performance of RDS estimators under varying conditions of trait prevalence, homophily, and relative activity. They used large simulated networks (N = 20,000) derived from real-world RDS degree reports and an empirical Facebook network (N = 22,470) to evaluate estimators of binary and categorical trait prevalence. Variability in prevalence estimates is higher when network degree is drawn from real-world samples than from the commonly assumed Poisson distribution, resulting in lower coverage rates. Newer estimators perform well when the sample is a substantive proportion of the population, but bias is present when the population size is unknown. The choice of preferred RDS estimator needs to be study specific, considering both statistical properties and knowledge of the population under study.

Correlation distances in social networks

Abstract In this work, we explore degree assortativity in complex networks, and extend its usual definition beyond that of nearest neighbours. We apply this definition to model networks, and describe a rewiring algorithm that induces assortativity. We compare these results to real networks. Social networks in particular tend to be assortatively mixed by degree in contrast to many other types of complex networks. However, we show here that these positive correlations diminish after one step and in most of the empirical networks analysed. Properties besides degree support this, such as the number of papers in scientific coauthorship networks, with no correlations beyond nearest neighbours. Beyond next-nearest neighbours we also observe a disassortative tendency for nodes three steps away indicating that nodes at that distance are more likely different than similar.

Generalized network density matrices for analysis of multiscale functional diversity.

The network density matrix formalism allows for describing the dynamics of information on top of complex structures and it has been successfully used to analyze, e.g., a system's robustness, perturbations, coarse-graining multilayer networks, characterization of emergent network states, and performing multiscale analysis. However, this framework is usually limited to diffusion dynamics on undirected networks. Here, to overcome some limitations, we propose an approach to derive density matrices based on dynamical systems and information theory, which allows for encapsulating a much wider range of linear and nonlinear dynamics and richer classes of structure, such as directed and signed ones. We use our framework to study the response to local stochastic perturbations of synthetic and empirical networks, including neural systems consisting of excitatory and inhibitory links and gene-regulatory interactions. Our findings demonstrate that topological complexity does not necessarily lead to functional diversity, i.e., the complex and heterogeneous response to stimuli or perturbations. Instead, functional diversity is a genuine emergent property which cannot be deduced from the knowledge of topological features such as heterogeneity, modularity, the presence of asymmetries, and dynamical properties of a system.

Signed networks - The role of negative links for the diffusion of innovation

We study the effects of negative links for the innovation diffusion process. To do so we employ an agent-based simulation model of a threshold diffusion process. Our results reveal that negative relationships can have a negative but also positive impact on the innovation diffusion. If network actors simply weigh the positive/negative signals coming from peers that have adopted the innovation, already few negative links will have a significant negative impact on the diffusion performance. If, however, network actors additionally consider the signals coming from peers that have not adopted the innovation, networks with a medium to high number of negative links create a ‘double-negative’ effect which drives the diffusion. Finally, an empirical network shows that our results hold also for real world network data. Our results emphasize the need to better understand how negative relationships, foe effects, and non-adopter effects can hinder or support the diffusion of innovations, potentially explaining why key innovations fail to spread and others succeed. The implications of this research are relevant for the discussion on the diffusion of innovation but also for the diffusion of information, opinions etc. It offers insights for researchers, companies, and policymakers wishing to scale and diffuse innovations more quickly.

Research on the health status evaluation method of rolling bearing based on EMD‐GA‐BP

AbstractTo more accurately evaluate the health state of rolling bearings, this paper proposes a health status evaluation method based on empirical pattern decomposition, genetic algorithm and BP neural network. Firstly, the vibration signal is decomposed by empirical mode decomposition (EMD) and the time domain features of each intrinsic mode function (IMF) component are extracted, and the signal‐to‐noise ratio (Snr) of the signal is improved effectively. Then, the initial threshold and weight of BP neural network are optimized by genetic algorithm, which effectively improves the Snr of the signal. Finally, the extracted features are input into the optimized BP neural network to realize the identification of different states of the bearing. The effectiveness of the method has been effectively verified in the bearing data of Case Western Reserve University bearing dataset and it has higher accuracy and robustness than other common evaluation methods.

Evading tipping points in socio-mutualistic networks via structure mediated optimal strategy

The threat of large-scale pollinator decline is increasing globally under stress from multiple anthropogenic pressures. Traditional approaches have focused on managing endangered species at an individual level, in which the effect of complex interactions such as mutualism and competition are amiss. Here, we develop a coupled socio-mutualistic network model that captures the change in pollinator dynamics with human conservation opinion in a deteriorating environment. We show that the application of social norm (or conservation) at the pollinator nodes is fit to prevent sudden community collapse in representative networks of varied topology. Whilst primitive strategies have focused on regulating abundance as a mitigation strategy, the role of network structure has been largely overlooked. Here, we develop a novel network structure-mediated conservation strategy to find the optimal set of nodes on which norm implementation successfully prevents community collapse. We find that networks of intermediate nestedness require conservation at a minimum number of nodes to prevent a community collapse. We claim the robustness of the optimal conservation strategy (OCS) after validation on several simulated and empirical networks of varied complexity against a broad range of system parameters. Dynamical analysis of the reduced model shows that incorporating social norms allows the pollinator abundance to grow that would have otherwise crossed a tipping point and undergo extinction. Together, this novel means OCS provides a potential plan of action for conserving plant–pollinator networks bridging the gap between research in mutualistic networks and conservation ecology.

Transitions and its indicators in mutualistic meta-networks: effects of network topology, size of metacommunities and species dispersal

Gradual changes in the environment could cause dynamical ecological networks to suddenly shift from one state to an alternative state. When this happens ecosystem functions and services provided by ecological networks get disrupted. We, however, know very little about how the topology of such interaction networks can play a role in the transition of ecological networks when spatial interactions come into play. In the event of such unwanted transitions, little is known about how statistical metrics used to inform such impending transitions, measured at the species-level or at the community-level could relate to network architecture and the size of the metacommunity. Here, using hundred and one empirical plant-pollinator networks in a spatial setting, I evaluated the impact of network topology and spatial scale of species interactions on transitions, and on statistical metrics used as predictors to forecast such transitions. Using generalized Lotka-Volterra equations in a meta-network framework, I show that species dispersal rate and the size of the metacommunity can impact when a transition can occur. In addition, forecasting such unwanted transitions of meta-networks using statistical metrics of instability was also consequently dependent on the topology of the network, species dispersal rate, and the size of the metacommunity. The results indicated that the plant-pollinator meta-networks that could exhibit stronger statistical signals before collapse than others were dependent on their network architecture and on the spatial scale of species interactions.

Identifying influential nodes by leveraging redundant ties

Structure-based influential nodes identification is a long-term challenge in the study of complex networks. While global centrality-based approaches are generally considered to be more accurate and reliable, the requirements of complete network information and high computational complexity are hard to meet, limiting their applications in many practical scenarios. In addition, recent studies have highlighted the effect of cyclic structures introducing redundant paths in network connectivity and exaggerating the importance of traditional centrality measures. In this work, we develop a new centrality metric, called Multi-Spanning Tree-based Degree Centrality (MSTDC), to quantify node importance with linear complexity by leveraging redundant ties. MSTDC is calculated using the aggregation of degrees of a small number of spanning trees constructed with a few randomly selected root nodes. Experiments on synthetic and empirical networks reveal that MSTDC obtains superior performance than other benchmark network centralities in identifying influential nodes from the perspective of both maintaining network connectivity and maximizing spreading capacity. In addition, we find that MSTDC is extraordinarily effective in networks with high clustering coefficients. Our study provides novel insights into the role of redundant ties in network structural and functional analyses.

Correlation dimension in empirical networks.

Network correlation dimension governs the distribution of network distance in terms of a power-law model and profoundly impacts both structural properties and dynamical processes. We develop new maximum likelihood methods which allow us robustly and objectively to identify network correlation dimension and a bounded interval of distances over which the model faithfully represents structure. We also compare the traditional practice of estimating correlation dimension by modeling as a power law the fraction of nodes within a distance to a proposed alternative of modeling as a power law the fraction of nodes at a distance. In addition, we illustrate a likelihood ratio technique for comparing the correlation dimension and small-world descriptions of network structure. Improvements from our innovations are demonstrated on a diverse selection of synthetic and empirical networks. We show that the network correlation dimension model accurately captures empirical network structure over neighborhoods of substantial size and span and outperforms the alternative small-world network scaling model. Our improved methods tend to lead to higher estimates of network correlation dimension, implying that prior studies could have produced or utilized systematic underestimates of dimension.

The duality of networks and groups: Models to generate two-mode networks from one-mode networks

AbstractShared memberships, social statuses, beliefs, and places can facilitate the formation of social ties. Two-mode projections provide a method for transforming two-mode data on individuals’ memberships in such groups into a one-mode network of their possible social ties. In this paper, I explore the opposite process: how social ties can facilitate the formation of groups, and how a two-mode network can be generated from a one-mode network. Drawing on theories of team formation, club joining, and organization recruitment, I propose three models that describe how such groups might emerge from the relationships in a social network. I show that these models can be used to generate two-mode networks that have characteristics commonly observed in empirical two-mode social networks and that they encode features of the one-mode networks from which they were generated. I conclude by discussing these models’ limitations and future directions for theory and methods concerning group formation.

Random Walk Sampling in Social Networks Involving Private Nodes

Analysis of social networks with limited data access is challenging for third parties. To address this challenge, a number of studies have developed algorithms that estimate properties of social networks via a simple random walk. However, most existing algorithms do not assume private nodes that do not publish their neighbors’ data when they are queried in empirical social networks. Here we propose a practical framework for estimating properties via random walk-based sampling in social networks involving private nodes. First, we develop a sampling algorithm by extending a simple random walk to the case of social networks involving private nodes. Then, we propose estimators with reduced biases induced by private nodes for the network size, average degree, and density of the node label. Our results show that the proposed estimators reduce biases induced by private nodes in the existing estimators by up to 92.6% on social network datasets involving private nodes.

Resolving mass transport losses at the catalyst layer of polymer electrolyte membrane fuel cells through semi-empirical modelling

In this study, the sources of oxygen transport resistance in a polymer electrolyte membrane (PEM) fuel cell are resolved using two different approaches: pore network modelling and empirical modelling. This is achieved by applying operando imaging and electrochemical performance testing results to the pore network model and empirical model, respectively. In the literature, the impact of liquid water on oxygen transport resistance in a 5-layer membrane electrode assembly is typically attributed to the accumulation of liquid water within the gas diffusion layer. A comparison of the oxygen transport resistance values predicted by pore network modelling to the values predicted by empirical modelling shows that empirical models significantly overpredict the contribution of the substrate within gas diffusion layers to oxygen transport resistance at limiting current conditions (up to 440%). Statistical analysis of the discrepancy (residual) between the values of the oxygen transport resistance of the substrate predicted by the empirical and pore network models indicates that the residual can be attributed to oxygen transport resistance in locations that are related to water production but independent of the relative humidity of the inlet: the catalyst layer (CL) and/or CL-microporous layer interface.

A fast algorithm to approximate the spectral density of locally tree-like networks with assortativity

AbstractGraphs have become crucial for representing and examining biological, social and technological interactions. In this context, the graph spectrum is an exciting feature to be studied because it encodes the structural and dynamic characteristics of the graph. Hence, it becomes essential to efficiently compute the graph’s spectral distribution (eigenvalue’s density function). Recently, some authors proposed degree-based methods to obtain the spectral density of locally tree-like networks in linear time. The bottleneck of their approach is that they assumed that the graph’s assortativity is zero. However, most real-world networks, such as social and biological networks, present assortativity. Consequently, their spectral density approximations may be inaccurate. Here, we propose a method that considers assortativity. Our algorithm’s time and space complexities are $\mathscr{O}(d_{\max}^{2})$, where $d_{\max}$ is the largest degree of the graph. Finally, we show our method’s efficacy in simulated and empirical networks.

Large network community detection by fast label propagation

Many networks exhibit some community structure. There exists a wide variety of approaches to detect communities in networks, each offering different interpretations and associated algorithms. For large networks, there is the additional requirement of speed. In this context, the so-called label propagation algorithm (LPA) was proposed, which runs in near-linear time. In partitions uncovered by LPA, each node is ensured to have most links to its assigned community. We here propose a fast variant of LPA (FLPA) that is based on processing a queue of nodes whose neighbourhood recently changed. We test FLPA exhaustively on benchmark networks and empirical networks, finding that it can run up to 700 times faster than LPA. In partitions found by FLPA, we prove that each node is again guaranteed to have most links to its assigned community. Our results show that FLPA is generally preferable to LPA.

Effects of adaptive cooperation among heterogeneous manufacturers on supply chain viability under fluctuating demand in post-COVID-19 era: an agent-based simulation

Supply chain viability (SCV) is associated with systematic and dynamic approaches to responding to COVID-19 disruptions. The supply chain has been disturbed due to the recurrent nature of the pandemic. In the post-pandemic era, large and small- and medium-sized manufacturers have both advantages and disadvantages, complementing each other. This study proposes an agent-based model to simulate the viable situation of heterogeneous supply systems by considering three behavioural rules (i.e. cooperation establishment, win-win cooperation, and cooperation priority). The results demonstrate that the viability will be improved if large manufacturers cooperate closely with small- and medium-sized manufacturers. The viability is also associated with two adjustment parameters of production capacity, indicating that large manufacturers need to become more scalable, flexible, and adaptable to survive a rapidly changing environment. In addition, when the cooperation is more widely-connected, the viability will be higher. When facing severe fluctuations, large and small- and medium-sized manufacturers can cooperate, acting as a ‘sponge system’ to flatten the fluctuations. Finally, the simulation results of an empirical network also support this conclusion. Therefore, we suggest that a viable supply chain design should consider integrating supply resources to improve the overall viability of the supply system.

The non-random assembly of network motifs in plant-pollinator networks.

Ecological processes leave distinct structural imprints on the species interactions that shape the topology of animal-plant mutualistic networks. Detecting how direct and indirect interactions between animals and plants are organised is not trivial since they go beyond pairwise interactions, but may get blurred when considering global network descriptors. Recent work has shown that the meso-scale, the intermediate level of network complexity between the species and the global network, can capture this important information. The meso-scale describes network subgraphs representing patterns of direct and indirect interactions between a small number of species, and when these network subgraphs differ statistically from a benchmark, they are often referred to as 'network motifs'. Although motifs can capture relevant ecological information of species interactions, they remain overlooked in natural plant-pollinator networks. By exploring 60 empirical plant-pollinator networks from 18 different studies with wide geographical coverage, we show that some network subgraphs are consistently under- or over-represented, suggesting the presence of worldwide network motifs in plant-pollinator networks. In addition, we found a higher proportion of densely connected network subgraphs that, based on previous findings, could reflect that species relative abundances are the main driver shaping the structure of the meso-scale on plant-pollinator communities. Moreover, we found that distinct subgraph positions describing species ecological roles (e.g. generalisation and number of indirect interactions) are occupied by different groups of animal and plant species representing their main life-history strategies (i.e. functional groups). For instance, we found that the functional group of 'bees' was over-represented in subgraph positions with a lower number of indirect interactions in contrast to the rest of floral visitors groups. Finally, we show that the observed functional group combinations within a subgraph cannot be retrieved from their expected probabilities (i.e. joint probability distributions), indicating that plant and floral visitor associations within subgraphs are not random either. Our results highlight the presence of common network motifs in plant-pollinator communities that are formed by a non-random association of plants and floral visitors functional groups.

Out of bounds? The boundary specification problem for centrality in psychological networks.

The analysis of psychological networks has become common in multiple subfields including clinical, social, and personality psychology, where the focus is often on identifying highly central nodes that represent symptoms, beliefs, or traits. However, the boundaries of these networks are often ambiguous and relevant nodes are often missing from the network. In this article, we use a series of simulations to show that even under typical conditions of missingness, the centrality of nodes in an empirical psychological network are poorly correlated or uncorrelated with their centrality in a hypothetical "true" psychological network, and thus are invalid. We illustrate the implications of this lack of validity using an empirical example drawn from a recent study of political belief system networks, demonstrating that the original study would have drawn incorrect conclusions about American's most central political beliefs. We conclude by recommending that centrality measures should be computed and interpreted only in psychological networks that include (nearly) all the nodes inside a theoretically meaningful boundary. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

A network of empirical ethics teams embedded in research programmes across multiple sites: opportunities and challenges in contributing to COVID-19 research and responses.

Covid-19 continues to teach the global community important lessons about preparedness for research and effective action to respond to emerging health threats. We share the COVID-19 experiences of a pre-existing cross-site ethics network-the Global Health Bioethics Network-which brings together researchers and practitioners from Africa, Europe, and Southeast Asia. We describe the network and its members and activities, and the work-related opportunities and challenges we faced over a one-year period during the pandemic. We highlight the value of having strong and long-term empirical ethics networks embedded across diverse research institutions to be able to: 1) identify and share relevant ethics challenges and research questions and ways of 'doing research'; 2) work with key stakeholders to identify appropriate ways to contribute to the emerging health issue response - e.g., through ethics oversight, community engagement, and advisory roles at different levels; and 3) learn from each other and from diverse contexts to advocate for positive change at multiple levels. It is our view that being embedded and long term offers opportunities in terms of deep institutional and contextual knowledge, existing relationships and access to a wide range of stakeholders. Being networked offers opportunities to draw upon a wide range of expertise and perspectives, and to bring together internal and external insights (i.e.drawing on different positionalities). Long term funding means that the people and resources are in place and ready to respond in a timely way. However, many tensions and challenges remain, including difficulties in negotiating power and politics in the roles that researchers and research institutions can and should play in an emergency, and the position of empirical ethics within research programmes. We discuss some of these tensions and challenges and consider the implications for our own and similar networks in future.