Community Detection Algorithm Research Articles

A Multi-Objective Pigeon-Inspired Optimization Algorithm for Community Detection in Complex Networks

Community structure is a very interesting attribute and feature in complex networks, which has attracted scholars’ attention and research on community detection. Many single-objective optimization algorithms have been migrated and modified to serve community detection problems. Due to the limitation of resolution, the final algorithm implementation effect is not ideal. In this paper, a multi-objective community detection method based on a pigeon-inspired optimization algorithm, MOPIO-Net, is proposed. Firstly, the PIO algorithm is discretized in terms of the solution space representation, position, and velocity-updating strategies to adapt to discrete community detection scenarios. Secondly, by minimizing the two objective functions of community score and community fitness at the same time, the community structure with a tight interior and sparse exterior is obtained. Finally, for the misclassification caused by boundary nodes, a mutation strategy is added to improve the accuracy of the final community recognition. Experiments on synthetic and real networks verify that the proposed algorithm is more accurate in community recognition compared to 11 benchmark algorithms, confirming the effectiveness of the proposed method.

Identifying beliefs about the gender gap in engineering professions among university students using community detection algorithms and statistical analysis

AbstractDigital societies require professionals in the Technology and Engineering sectors, but their lack, particularly of women, requires a thorough understanding of this gender gap. This research analyzes the beliefs and opinions of university engineering students about the gender gap in their professional fields by means of a community detection algorithm to identify groups of students with similar belief patterns. This study leverages a community detection algorithm to analyze the beliefs of 590 engineering students regarding the gender gap in their field, together with a correlational and explanatory design using a quantitative paradigm. A validated questionnaire focusing on the professional dimension was used. The algorithm identified three student communities, two gender‐sensitive and one gender‐insensitive. The study uncovered a concerning lack of awareness regarding the gender gap among engineering students. Many participants did not recognize the importance of increasing the representation of professional women, maintained the belief that the gender gap affects only women, and assumed that men and women are equally paid. However, women show a higher level of awareness, while men perceive the gender gap as a passing trend, which is worrying. Students recognize the importance of integrating a gender perspective into university and engineering curricula. It is worrying that many students doubt the existence of the gender gap and that both genders lack knowledge about gender gap issues. Finally, community detection algorithms could efficiently and automatically analyze gender gap issues or other unrelated topics.

Analysis of professional competency awareness based on visible network graphs

Introduction: With the advancement of society and technology, the importance of professional competency has been increasingly emphasized. This study aims to provide an in-depth analysis of the Baidu search index for “professional competency” from July 3, 2017, to October 6, 2023, across 34 provincial-level administrative regions in China.Methods: The study begins with a literature review to establish a theoretical framework. Subsequently, it delves into the fundamental concepts of complex networks, their parameters, and the principles of visible network graph algorithms. The application of system clustering and community detection algorithms is explored within this framework.Results: The empirical analysis starts with a provincial-level examination, involving the construction of visible network graphs, the exploration of scale-free networks, and discussions on regional system clustering. Following this, a nationwide analysis is conducted, focusing on the construction of visible network graphs to visualize the awareness of professional competency and studying community structures.Discussion: The findings provide insights into the awareness of professional competency across different regions. These insights are valuable for informing policy formulation and corporate decision-making, contributing to a deeper understanding of professional competency awareness and its implications.

Comparing the Clique Percolation algorithm to other overlapping community detection algorithms in psychological networks: A Monte Carlo simulation study.

In psychological networks, one limitation of the most used community detection algorithms is that they can only assign each node (symptom) to a unique community, without being able to identify overlapping symptoms. The clique percolation (CP) is an algorithm that identifies overlapping symptoms but its performance has not been evaluated in psychological networks. In this study, we compare the CP with model parameters chosen based on fuzzy modularity (CPMod) with two other alternatives, the ratio of the two largest communities (CPRat), and entropy (CPEnt). We evaluate their performance to: (1) identify the correct number of latent factors (i.e., communities); and (2) identify the observed variables with substantive (and equally sized) cross-loadings (i.e., overlapping symptoms). We carried out simulations under 972 conditions (3x2x2x3x3x3x3): (1) data categories (continuous, polytomous and dichotomous); (2) number of factors (two and four); (3) number of observed variables per factor (four and eight); (4) factor correlations (0.0, 0.5, and 0.7); (5) size of primary factor loadings (0.40, 0.55, and 0.70); (6) proportion of observed variables with substantive cross-loadings (0.0%, 12.5%, and 25.0%); and (7) sample size (300, 500, and 1000). Performance was evaluated through the Omega index, Mean Bias Error (MBE), Mean Absolute Error (MAE), sensitivity, specificity, and mean number of isolated nodes. We also evaluated two other methods, Exploratory Factor Analysis and the Walktrap algorithm modified to consider overlap (EFA-Ov and Walk-Ov, respectively). The Walk-Ov displayed the best performance across most conditions and is the recommended option to identify communities with overlapping symptoms in psychological networks.

Leveraging cascading information for community detection in social networks

Information exchange among individuals is one of the key factors leading to formation of modular structures in social networks, often referred as communities. In this paper, we devise a novel information diffusion based approach to leverage the latent information exchange among the individuals for identifying communities. Designed an algorithm called PraSar, which has mainly two phases namely seeding and unification. In seeding phase, closely connected groups of nodes called seed communities are identified by analyzing the cascades obtained during information diffusion. In unification phase, the external connections of seed communities are reduced by unifying two or more seed communities. Involvement of cascade that are formed during the diffusion process enables PraSar to operate locally and giving an overall complexity of O(n2). The theoretical foundation of the proposed approach is established by various theorems. Empirical results on diverse real world datasets are evident for the effectiveness and competitiveness of the PraSar algorithm over state-of-the-art community detection algorithms.

Urban mobility analytics amid COVID-19 pandemic: A framework for promoting work resumption based on mobile phone data

The Corona Virus Disease 2019 (COVID-19) pandemic had a pernicious influence on the whole world, so that the international community implemented various travel policies to inhibit the viral spread. With the COVID-19 pandemic under control, the overly lenient measures allowed the new variant to take advantage. Also, it is difficult to determine the geographical scope of transport policy that has the least impact on work resumption. Therefore, it is necessary to find policies to promote work resumption while preventing and controlling the pandemic in a geographical perspective. In this paper, a framework based on mobile phone data is proposed for promoting work resumption during the pandemic. First, human mobility networks are constructed through the origin-destination flow extracted from mobile phone data. Second, the singular value decomposition (SVD) algorithm is utilized to recognize the mobility patterns and explore the urban mobility dynamics of work-resumption-related travel during the pandemic. Third, the nonoverlapping community detection algorithm is used to generate policy zoning for work-resumption-related travel. Finally, the policy zoning is validated by public health data based on the overlay analysis and spatial statistics method, and policy implications are proposed. A case study of Shanghai is applied to verify the proposed framework. The results show that the framework can capture the mobility patterns and implement the policy zoning from mobile phone data to balance the economic activity and public health during the work resumption, and the proposed framework can be applied to other cities with future public health events.

Community hiding: Completely escape from community detection

Community detection algorithms play an important role in revealing and analyzing complex network structures. However, when these algorithms are put into practice, privacy concerns often arise, especially when community nodes wish to safeguard their identities from exposure. This interaction between network analysis and privacy issues leads to what we refer to as the “community hiding problem.” In the context of concealing specific target communities, popular methods primarily focus on preventing community detection algorithms from accurately identifying the entire original community, resulting in partial concealment. This study formally addresses the challenge of achieving complete concealment of target communities. We introduce key metrics such as escape scores, dispersion scores, and hiding scores to precisely define the problem. Additionally, we design a metric, denoted as M-value, to evaluate the effectiveness of concealing target communities from multiple perspectives. To tackle this challenge, we employ a genetic algorithm that leverages previous knowledge to maximize the concealment of the target community while minimizing changes to link connections. We validate our approach through extensive experiments with various real-world datasets, demonstrating that our algorithm outperforms several state-of-the-art baseline algorithms across multiple metrics. Visual representations of our method confirm its ability to effectively hide target communities while minimally affecting the broader network's community structure, strengthening the inherent effectiveness of our community hiding approach. Furthermore, we assess the adaptability of our modified network with different community detection algorithms, consistently demonstrating effective concealment even when these algorithms are extended. This emphasizes the robustness and generality of our proposed approach in various algorithmic scenarios.

Is corrections officers' use of illegal force networked? Network structure, brokerage, and key players in the New York City Department of Correction

PurposeRecent research on police misconduct has adopted a network perspective, while recent work on correctional officer (CO) use of force has similarly framed it as an organizational behavior, learned through socialization and deployed with considerable discretion. This paper leverages the network paradigm in police misconduct research to study the use of force by COs. MethodsUsing data on lawsuits involving the New York City Department of Correction from 2013 to 2022, this paper forms a co-offending network of those COs who have been sued for excessive force. This paper then uses descriptive social network analysis and key player selection and community detection methods. ResultsAfter describing the basic network structure of COs who use force together, this paper identifies brokers who connect disparate parts of the CO network; it also uses a community detection algorithm to identify clusters of COs involved in force incidents. Finally, the paper compares this network to an analogous network of police use of force, finding structural similarities. ConclusionsThis analysis bridges disparate work on correctional officer use of force and police misconduct; for policymakers who seek to curtail COs' excessive use of force, this paper might provide a blueprint for identifying key players and clusters.

Multivariate time series clustering based on fuzzy cognitive maps and community detection

Most time series clustering methods mainly focus on univariate time series (UTS). Compared with UTS, multivariate time series (MTS) consists of multiple components. Although interest in MTS clustering is increasing, its performance is far from satisfactory. Most traditional MTS clustering methods may have two limitations. First, they do not consider both the temporal features of each component and the relationship between the components. Second, they can only identify the local relationship between adjacent data, but cannot obtain long-distance global relationships and capture clusters of arbitrary shapes. In this paper, we develop a method for MTS clustering based on fuzzy cognitive maps (FCMs) and community detection, termed as MTSC-FCM-CD. To overcome the first limitation, we use FCM to represent MTS; FCM can extract temporal features while preserving the relationship between components in MTS. To overcome the second limitation, we use the community detection algorithm to cluster the global relations, which is different from the traditional nearest neighbor distance-based method. In the calculation process, the similarity between FCMs should be used to build a complex network. Existing methods calculate the similarity between FCMs, only considering the numerical characteristics but ignoring the topological structure. To solve this problem, we design a two-stage similarity measure to build a complex network. In comparison to the existing methods, the experimental results on twenty benchmark datasets demonstrate the effectiveness of MTSC-FCM-CD in MTS clustering.

The equivalences of community detection methods for bipartite networks

Abstract Investigating the community structures of bipartite networks is a frequent topic of discussion in the ecological and social fields. The most widely used methods, as proposed by numerous academics from varying perspectives, include spectral graph partitioning, modularity, nonnegative matrix factorization, and stochastic block model. In this paper, we demonstrate three equivalences among these four methods. One, both Dhillon spectral graph partitioning and Barber modularity clustering are equivalent to solving for the matrix's left and right singular vectors after relaxing the discrete constraints. Two, the nonnegative matrix factorization clustering is equivalent to the Dhillon spectral graph partitioning. Three, The bipartite stochastic block model is equivalent to the constraint-based NMF that uses K-L divergence as its cost function. These equivalences, obtained through rigorous mathematical derivations, will aid in the future development of efficient algorithms for community detection in bipartite networks.

Link Prediction and Graph Structure Estimation for Community Detection

In real-world scenarios, obtaining the relationships between nodes is often challenging, resulting in incomplete network topology. This limitation significantly reduces the applicability of community detection methods, particularly neighborhood aggregation-based approaches, on structurally incomplete networks. Therefore, in this situation, it is crucial to obtain meaningful community information from the limited network structure. To address this challenge, the LPGSE algorithm was designed and implemented, which includes four parts: link prediction, structure observation, network estimation, and community partitioning. LPGSE demonstrated its performance in community detection in structurally incomplete networks with 10% missing edges on multiple datasets. Compared with traditional community detection algorithms, LPGSE achieved improvements in NMI and ARI metrics of 1.5781% to 29.0780% and 0.4332% to 31.9820%, respectively. Compared with similar community detection algorithms for structurally incomplete networks, LPGSE also outperformed other algorithms on all datasets. In addition, different edge-missing ratio settings were also attempted, and the performance of different algorithms in these situations was compared and analyzed. The results showed that the algorithm can still maintain high accuracy and stability in community detection across different edge-missing ratios.

Proof of biased behavior of Normalized Mutual Information

The Normalized Mutual Information (NMI) metric is widely utilized in the evaluation of clustering and community detection algorithms. This study explores the performance of NMI, specifically examining its performance in relation to the quantity of communities, and uncovers a significant drawback associated with the metric's behavior as the number of communities increases. Our findings reveal a pronounced bias in the NMI as the number of communities escalates. While previous studies have noted this biased behavior, they have not provided a formal proof and have not addressed the causation of this problem, leaving a gap in the existing literature. In this study, we fill this gap by employing a mathematical approach to formally demonstrate why NMI exhibits biased behavior, thereby establishing its unsuitability as a metric for evaluating clustering and community detection algorithms. Crucially, our study exposes the vulnerability of entropy-based metrics that employ logarithmic functions to similar bias.

Motif-based community detection in heterogeneous multilayer networks

Multilayer networks composed of intralayer edges and interlayer edges are an important type of complex networks. Considering the heterogeneity of nodes and edges, it is necessary to design more reasonable and diverse community detection methods for multilayer networks. Existing research on community detection in multilayer networks mainly focuses on multiplexing networks (where the nodes are homogeneous and the edges are heterogeneous), but few studies have focused on heterogeneous multilayer networks where both nodes and edges represent different semantics. In this paper, we studied community detection on heterogeneous multilayer networks and proposed a motif-based detection algorithm. First, the communities and motifs of multilayer networks are defined, especially the interlayer motifs. Then, the modularity of multilayer networks based on these motifs is designed, and the community structure of the multilayer network is detected by maximizing the modularity of multilayer networks. Finally, we verify the effectiveness of the detection algorithm on synthetic networks. In the experiments on synthetic networks, comparing with the classical community detection algorithms (without considering interlayer heterogeneity), the motif-based modularity community detection algorithm can obtain better results under different evaluation indexes, and we found that there exists a certain relationship between motifs and communities. In addition, the proposed algorithm is applied in the empirical network, which shows its practicability in the real world. This study provides a solution for the investigation of heterogeneous information in multilayer networks.

A hybrid information-based two-phase expansion algorithm for community detection with imbalanced scales

A hybrid information-based two-phase expansion algorithm for community detection with imbalanced scales

Identifying Communities with Modularity Metric Using Louvain and Leiden Algorithms

Over the past 20 years, there has been a significant increase in publication in complex network analysis research, especially in community detection. Many methods were proposed to identify community structure. Each community identification algorithm has strengths and weaknesses due to the complexity of information. Among them, the optimisation methods are widely focused on. This paper focuses on an empirical study of two community detection algorithms based on agglomerative techniques using modularity metric: Louvain and Leiden. In this regard, the Louvain algorithm has been shown to produce a bad connection in the community and disconnected when executed iteratively. Therefore, the Leiden algorithm is designed to successively resolve the weaknesses. Performance comparisons between the two and their concept were summarised in detail, as well as the step-by-step learning process of the state-of-the-art algorithms. This study is important and beneficial to the future study of interdisciplinary data sciences of network analysis. First, it demonstrates that the Leiden method outperformed the Louvain algorithm in terms of modularity metric and running time. Second, the paper displays the use of these two algorithms on synthetic and real networks. The experiment was successful as it identified better performance, and future work is required to confirm and validate these findings.

Community Deception in Large Networks: Through the Lens of Laplacian Spectrum

Many complex networks in the real world have community structures. Typical examples include online social networks and ecology networks. While the identification of communities bears numerous practical applications, with the increasing awareness of data security and privacy concerns, the need to protect the community affiliations of individuals from disclosing by attackers emerges. This raises the community deception (CD) problem, that is, the opposite of community detection, which asks for ways to minimally perturb the network structures by rewiring nodes so that the target communities maximally hide themself from community detection algorithms. To this end, we investigate the CD problem through a Laplacian spectrum lens and propose a method named <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mathtt{ComDeceptor}$</tex-math> </inline-formula> to hide a flexible target set of communities, which is more universal than most existing methods that either focus on hiding the entire communities or a single community. The key idea of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mathtt{ComDeceptor}$</tex-math> </inline-formula> is to first allocate the resources of perturbations fairly and effectively. By proving that hiding communities through intercommunity edge addition and intracommunity edge deletion correspond to maximizing the second smallest eigenvalue <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\lambda_2$</tex-math> </inline-formula> and minimizing the largest eigenvalue <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\lambda_n$</tex-math> </inline-formula> of the graph Laplacian, respectively, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mathtt{ComDeceptor}$</tex-math> </inline-formula> then incorporates efficient heuristics for approximately solving the problems, thus selecting the appropriate edge to perturb. Experimental results over nine real-world networks and six community detection algorithms not only demonstrate the efficiency of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mathtt{ComDeceptor}$</tex-math> </inline-formula> , but also the superior performance on obfuscating community structures over the baselines.

Spatial connectivity of reef manta rays across the Raja Ampat archipelago, Indonesia.

The reef manta ray Mobula alfredi is present throughout most island groups that form the Raja Ampat archipelago, Indonesia. The species is protected regionally and nationally and is currently managed as a single homogeneous population within the 6.7 million ha archipelago. However, scientific evidence is currently lacking regarding the spatial connectivity and population structure of M. alfredi within this archipelago. Using network analysis and an array of 34 acoustic receivers deployed throughout Raja Ampat between February 2016 and September 2021, we examined the movements of 72 subadult and adult M. alfredi tagged in seven regions of Raja Ampat. A total of 1094 M. alfredi movements were recorded and were primarily concentrated between nearby receiver stations, highlighting frequent local movements within, and limited long-distance movements between regional acoustic receiver arrays. Network analysis revealed highly connected nodes acting as hubs important for M. alfredi movements. A community detection algorithm further indicated clusters within the network. Our results suggest the existence of a metapopulation comprising three demographically and geographically distinct subpopulations within the archipelago. They also reveal the importance of Eagle Rock as a critical node in the M. alfredi movement network, justifying the urgent inclusion of this site within the Raja Ampat marine protected area network.

Community detection in hypergraphs via mutual information maximization

The hypergraph community detection problem seeks to identify groups of related vertices in hypergraph data. We propose an information-theoretic hypergraph community detection algorithm which compresses the observed data in terms of community labels and community-edge intersections. This algorithm can also be viewed as maximum-likelihood inference in a degree-corrected microcanonical stochastic blockmodel. We perform the compression/inference step via simulated annealing. Unlike several recent algorithms based on canonical models, our microcanonical algorithm does not require inference of statistical parameters such as vertex degrees or pairwise group connection rates. Through synthetic experiments, we find that our algorithm succeeds down to recently-conjectured thresholds for sparse random hypergraphs. We also find competitive performance in cluster recovery tasks on several hypergraph data sets.

Developing a novel approach in estimating urban commute traffic by integrating community detection and hypergraph representation learning

The efficiency of urban traffic management and congestion alleviation relies heavily on accurate forecasting of Origin-Destination (O-D) demand matrices. Existing models primarily focus on estimating O-D demand for various travel purposes throughout the day, which is characterised by its pulsating nature. However, these models often compromise the precision of peak-hour forecasts, leading to unreliable dynamic traffic control and challenges in effectively reducing peak-hour congestion. To tackle this challenge, this paper proposes a novel method for predicting commuting O-D demand matrices. Our method employs community detection algorithms on road networks to precisely partition commute O-D regions, incorporating Points of Interest (POIs). We also present a spatio-temporal dynamic weighted hypergraph model that leverages these partitioned regions, time characteristics from observed O-D trips, and meteorological data to improve forecasting. Comparative analyses with contemporary models and ablation studies indicate our method significantly enhances prediction accuracy, by approximately 5%. These findings imply that the proposed method more effectively encompasses the varied characteristics of commuting during peak hours, thereby providing more accurate demand matrices for urban traffic management.

Exploring Systemic Risk Dynamics in the Chinese Stock Market: A Network Analysis with Risk Transmission Index

Systemic risk refers to the potential for a disruption in one part of a financial system to trigger a cascade of adverse effects, impacting the functioning of the system. Despite the progress on novel systemic risk measures, research on dynamics of systemic risk network structure and its community effect is still in its initial state. In this study, we utilize price data from 107 representative Chinese stocks spanning the period from 2017 to 2022. A systemic risk network is derived from the Risk Transmission Index based on TENET and the QR–Lasso model. By utilizing DBSCAN, HITS and community detection algorithms on the network, we aim to propose a more suitable definition of systemically important companies, explore the interrelationships between companies, and discuss its plausible reasons for dynamics structural changes. The empirical findings demonstrate a substantial involvement of insurance companies in both contributing to and receiving systemic risk within the analyzed context. We identify prominent risk output and input centers, and emphasize the profound impact of the COVID-19 pandemic on the dynamics of systemic risk.