An EDA-based community detection in complex networks

There is an increasing recognition on community detection in complex networks in recent years. In this study, we improve a recently proposed memetic algorithm for community detection in networks. By introducing a Population Generation via Label Propagation (PGLP) tactic, an Elitism Strategy (ES) and an Improved Simulated Annealing Combined Local Search (ISACLS) strategy, the improved memetic algorithm called (iMeme-Net) is put forward for solving community detection problems. Experiments on both computer-generated and real-world networks show the effectiveness and the multi-resolution ability of the proposed method.

A multi-objective adaptive evolutionary algorithm to extract communities in networks

The problem of community detection is relevant in many disciplines of science and modularity optimization is the widely accepted method for this purpose. It has recently been shown that this approach presents a resolution limit by which it is not possible to detect communities with sizes smaller than a threshold, which depends on the network size. Moreover, it might happen that the communities resulting from such an approach do not satisfy the usual qualitative definition of commune; i.e., nodes in a commune are more connected among themselves than to nodes outside the commune. In this paper we present a different method for community detection in complex networks. We define merit factors based on the weak and strong community definitions formulated by Radicchi [Proc. Natl. Acad. Sci. U.S.A. 101, 2658 (2004)] and we show that these local definitions avoid the resolution limit problem found in the modularity optimization approach.

Multiplex community detection in complex networks using an evolutionary approach

Community detection is an important task in the analysis of complex networks, which is significant for mining and analyzing the organization and function of networks. As an unsupervised learning algorithm based on the particle competition mechanism, stochastic competitive learning has been applied in the field of community detection in complex networks, but still has several limitations. In order to improve the stability and accuracy of stochastic competitive learning and solve the problem of community detection, we propose an unsupervised community detection algorithm LNSSCL (Local Node Similarity-Integrated Stochastic Competitive Learning). The algorithm calculates node degree as well as Salton similarity metrics to determine the starting position of particle walk; local node similarity is incorporated into the particle preferential walk rule; the particle is dynamically adjusted to control capability increments according to the control range; particles select the node with the strongest control capability within the node to be resurrected; and the LNSSCL algorithm introduces a node affiliation selection step to adjust the node community labels. Experimental comparisons with 12 representative community detection algorithms on real network datasets and synthetic networks show that the LNSSCL algorithm is overall better than other compared algorithms in terms of standardized mutual information (NMI) and modularity (Q). The improvement effect for the stochastic competition learning algorithm is evident, and it can effectively accomplish the community detection task in complex networks.

Community detection in complex networks is an important multidisciplinary research area and is considered crucial for understanding the structure of complex networks. Unsupervised deep learning models (e.g. stack autoencoders) have been successfully proposed for the problem of community detection, which can extract network features and use them in splitting the network into communities. Despite their effectiveness, these methods are not very efficient, especially when large networks are involved. Furthermore, existing models usually handle the network as a single object, which means that massive trainable parameters are required during training (which in normal complex networks often reach millions of parameters, not to mention large networks), thereby increasing the complexity of the model. To overcome these problems, this paper proposes a deep autoencoder method for network community detection based on three techniques: network-data partitioning, reduction and sharing of trainable parameters, which contribute significantly to improve the efficiency of the method. A new partitioning strategy is imposed on complex networks at different levels. The paper also proposes a parallel design for the proposed method. Furthermore, a new similarity constraint function is proposed to improve and preserve the performance of community detection task. We performed extensive experiments for different partitioning levels of a network-dataset to evaluate the method with CPU and GPU devices. The results showed that the proposed method significantly improved training speed and efficiency while maintaining performance. The results also showed that the efficiency of the method increases as we move to a deeper level of partitioning.

Complex networks (CNs) have gained much attention in recent years due to their importance and popularity. The rapid growth in the size of CNs leads to more difficulties in the analysis of CNs tasks. Community Detection (CD) is an important multidisciplinary research area where many machine/deep learning-based methods have been applied to map CNs into a low-dimensional representation for extracting information similarity among members of CNs. Currently, Deep Learning (DL) is one of the promising methods to extract knowledge and learn information from high dimensional space and represent it in low dimensional space. However, designing an accurate and efficient DL-based CD method especially when dealing with large CNs is always an on-going research endeavor to pursue. Meta-Heuristic (MH) algorithms have shown their potentials in improving DL models in terms of solution quality and computational cost. In addition, parallel computing is a feasible solution for building efficient DL models. The algorithmic principle of MH is parallel in nature; however, its computation framework in DL training that is reported in the literature is not really implemented in a parallel computing setup. In this paper, we present a systematic review of CD in CNs from conventional machine learning to DL methods and point out the gap of applying DL-based CD methods in large CNs. In addition, the relevant studies on DL with parallel and MH approaches are reviewed and their implications on DL models are highlighted to prospect effective solutions to overcome the challenges of DL-based CD methods. We also point out research challenges in the field of CD and suggest possible future research directions.

Detecting communities in complex networks and particularly in social networks have been gaining traction in recent years. Not much work has been undertaken in the area of Community Detection for patient-centric social networks. In this work, we have tried to detect significant communities in a patient-centric social network by computing similarity among the unstructured physician notes which are in the form of long texts. Experiments have been conducted using the MIMIC III dataset, which is a large open database containing de-identified health-related patient data. Three modularity based community detection algorithms, viz., FastGreedy, Walktrap, and Louvain have been applied to identify the potential communities in the network. Further, the significance of such communities has been tested using the Wilcoxon rank-total test. It is found that Louvain algorithm fared better in terms of modularity and significance compared to the other two algorithms.

Community structure in networks indicates some meaningful groups or organizations in the real world. Various algorithms to detect community in complex networks were proposed. However, problems about how to judge the goodness and performance of algorithms are still open. This paper reviewed and analyzed the related work and algorithms for community detection, hoping to benefit researchers in related field.

<p dir="ltr">One of the most critical features of real-world networks is their community structure, as it reveals the internal organization of nodes. Finding such partitions is an interesting and essential task from both theoretical and practical perspectives. As a result, there is a growing demand for effective community detection algorithms. However, testing and designing such algorithms is demanding and challenging: one of the biggest problems is the lack of proper datasets and the fact that some have disputable ground-truth labeling, to say the least. One way to avoid this problem is to use artificial benchmarks with a built-in community structure.</p><p dir="ltr">In this work, we propose ABCDe—a multi-threaded implementation of the ABCD (Artificial Benchmark for Community Detection) graph generator. We test its performance and compare it with the previously available sequential version of the ABCD model and the parallel implementation of the standard and widely used LFR generator. We also investigate various theoretical asymptotic properties of the ABCD and ABCDe models. In particular, we analyze the modularity function, arguably the most important graph property in the context of community detection. By doing so, we show that models from the ABCD family are proper tools for designing new algorithms, not only those suited specifically for the community detection task.</p><p dir="ltr">Finally, we introduce a new approach to community detection in graphs. The idea is to initialize one of the popular heuristic-based community detection algorithms, such as Louvain or Leiden, with a partition obtained by embedding the graph nodes into a high-dimensional real vector space and then running a clustering algorithm on this latent representation. This procedure significantly improves the results and reduces the inherent volatility of such algorithms. Proper embedding filters out unnecessary information while retaining the proximity of nodes belonging to the same community.</p>

Previous studies on network mining have focused primarily on learning a single task (such as classification or community detection) on a given network. This paper considers the problem of multi-task learning on heterogeneous network data. Specifically, we present a novel framework that enables one to perform classification on one network and community detection in another related network. Multi-task learning is accomplished by introducing a joint objective function that must be optimized to ensure the classes in one network are consistent with the link structure, nodal attributes, as well as the communities detected in another network. We provide both theoretical and empirical analysis of the framework. We also show that the framework can be extended to incorporate prior information about the correspondences between the clusters and classes in different networks. Experiments performed on both real-world and synthetic data sets demonstrate the effectiveness of the joint framework compared to applying classification and community detection algorithms on each network separately.

Community detection in complex networks has increasingly become an important topic in the network, but in most community detection methods, a single node only belongs to one community. In fact, there is often overlap among real-world online communities. In this paper, an overlapping community detection algorithm based on subgraph structure and multi-optimization method is designed. In this algorithm, the maximum clique mined by k-core decomposition is used as the clique node, thus the overlap characteristic is transformed into the inherent characteristic of the new graph. After that, a population initialization method based on k-core decomposition is designed, and the discrete framework of multi-objective particle swarm optimization algorithm is used to optimize the two objectives on the basis of maximal clique graph to solve the problem of overlapping community detection. In the real-world network, this algorithm is compared with similar community detection algorithms. The comparison of the evaluation indexe shows that the community detection effect of this algorithm is similar to that of similar algorithms, and has a good application prospect.

Communities are virtually ubiquitous in real-world networks, and the statistic of modularity index Q is the classical measurement for community detection algorithms. However, the relationship between the modularity property and network multilever micro-scale structures is still not clear. In this paper, we study community detection results both in artificial and real-life complex networks by constructing different order null networks, and the results uncover that how micro-structures (such as degree distribution, assortativity and clustering coefficient) affect community properties. Meanwhile, we also propose two novel null networks (increasing or decreasing community structures) to verify the robustness of different community detection algorithms. Our results indicate that the modularity index Q is not a suitable statistic to measure the weak community property which is widely available in empirical networks. Our findings can not only be used to test the robustness of different community detection methods, but also be helpful to uncover the correlation of network structures between microcosmic and mesoscopic scales.

Community detection in complex networks via adapted Kuramoto dynamics

Complex networks represent various real‐world systems. Overlapping community detection is one of the critical tasks in studying these networks and has significance to a wide variety of applications, including the exploration of online social networks because of the natural attitude of persons to participate in multiple communities at the same time. Despite a large number of existing community detection algorithms for detecting disjoint communities, the efficient and fast uncovering of overlapping communities has remained a challenging problem. To provide an efficient solution, on the one hand, the balanced link density label propagation (BLDLP) algorithm, proposed by the authors of the current study, is a fast, stable, and efficient method for disjoint community detection. On the other hand, the fuzzy theory is a worthwhile approach for overlapping community detection since it provides the membership rate of the overlapping nodes as well as the detection of overlapping communities. Hence, in this paper, based on the synergy of the BLDLP algorithm and the fuzzy theory, a novel method, called fuzzy BLDLP, for overlapping community detection is proposed. Fuzzy BLDLP is fast and efficient. The proposed method needs no prior information about the number of network communities to discover them. The experiments on both synthetic and real‐world known networks, including Zachary, Dolphins, and COVID‐19 Co‐authorship, have revealed that the proposed method successfully detects the overlapping nodes and communities and hence is comparable with the state‐of‐the‐art overlapping community detection algorithms in terms of recall, precision, F‐score and overlapping normalized mutual information.