A Survey of Parallel Community Detection Algorithms

As the online social network (OSN) develops, the social media data can be easily obtained, which leads to a large scale social network data. In our offline social networks, social relationships that an individual (ego) maintains with other people (alters) can also be organized into circles or groups according to the ego network model [1]. As to online social network, in order to help users cope with personal social network, many OSN sites provide services for users to group their friends. With many restrictions, such as on Facebook one cannot access other's personal information unless they establish ties with each other, the network structure is the only information known by users. In this paper, we propose an algorithm to find circles on personal social network mainly from the structural view. Based on the defined structural similarity and the condition where nodes constitute a circle, the algorithm is able to cluster the most similar nodes into the same circle. We then give proof to demonstrate the convergence of our algorithm. To show the effectiveness of our algorithm, we realize the algorithm on Facebook datasets. Compared with the community detection algorithm, the results are evaluated from the visualized quantified aspect. From the results, we find that community detection algorithm cannot directly deal with the circle detection problem due to the characteristic of ego network and our algorithm can effectively handle the circle detection on ego network.

Community detection is a well-known area of research that yields essential results in various fields such as social media, biological networks, pandemic spread, recommendation systems, etc. There are two important areas for improvement in existing community detection algorithms: quality of community formed should be improved and a parallel approach to community detection is needed to handle massive data. In this paper, we have proposed a three-step parallel algorithm, Par-Com, using the concept of k Clique and modularity optimization to address both of the above issues. The proposed algorithm increases the execution speed and also improves the quality of the community formed by optimizing modularity. Par-Com uses dynamic load balancing on the multicore architecture of Supercomputer ParamShivay. We have also evaluated Par-Com’s performance against nine sequential and three parallel community detection algorithms on varying size datasets, i.e. karate, macaque, email, immuno, soc-epinions, facebook, and com-Friendster. The experiment result shows that Par-Com outperforms other algorithms under consideration with up to 45% increase in modularity and up to 84% increase in execution speed. Par-Com is also capable of detecting overlapping communities, fuzzy membership of each node, most influential node in each community formed, and outlier nodes. Nodes within a community that have the highest influence are deemed as experts. The choices made by an expert in a particular community are served as recommendations to other users within that community.

Two birds with one stone: Classifying positive and unlabeled examples on uncertain data streams

SUMMARYWith the rapid progress of the Internet, more and more people socialize and make new friends through online social network sites and applications, such as Facebook, Twitter, and MSN Messenger. The number of users of these online social network sites and applications has increased significantly within a short period. Unfortunately, online social networks are often the platforms that propagate malicious software, such as viruses. The malicious software is spread by infected users who automatically send fake requests to other users. After accepting the fake requests, the users are infected. In a realistic environment, everyone can decide whether to accept or decline a request, and thus, we study how to construct and use community structures to efficiently control virus propagation in online social networks whose users each have a probability, namely, inclination, of accepting a request. In this paper, a community detection algorithm is proposed to detect communities in online social networks. In addition, a number of users are selected from the communities for patching and controlling virus propagation. Simulation results show that our proposed method provides good performance in terms of the number of distributed patches.Copyright © 2013 John Wiley & Sons, Ltd.

Community detection is a dominant research field in social media mining. Numerous algorithms have been developed to detect community structure in network. Each of them focuses on different characteristic of data and generates different communities for same dataset. So cluster validation or comparative analysis of these algorithms is required to compare the efficiency of these algorithms. There are two known ways for cluster validation: Internal Quality Measure (IQM) and External Quality Measure (EQM). IQM evaluates the cluster without reference to external information while EQM requires already known clusters i.e. ground truth for comparison. Most of the researchers have compared Community Detection algorithms on the basis of either IQM or EQM but not both. In this paper, we have tried to compare nine existing sequential community detection algorithms on the basis of both IQM and EQM and we observed that overall performance of Louvain, Fastgreedy and Walktrap algorithm is more efficient than other algorithms. From this paper, researchers working in this area can get information about major sequential community detection algorithms and their efficiency on small real world datasets. They can incorporate the findings to develop more efficient algorithms.

Given a set A of activities expressed via temporal stochastic automata, and a set O of observations (detections of low level events), we study the problem of identifying instances of activities from A in O. While past work has developed algorithms to solve this problem, in this paper, we develop methods to significantly scale these algorithms. Our PASS architecture consists of three parts: (i) leveraging past work to represent all activities in A via a single “merged” graph, (ii) partitioning the graph into a set of C subgraphs, where (C + 1) is the number of compute nodes in a cluster, and (iii) developing a parallel activity detection algorithm that uses a different compute node in the cluster to intensively process each subgraph. We propose three possible partitioning methods and a parallel activity-search detection (PASS_Detect) algorithm that coordinates computations across nodes in the cluster. We report on experiments showing that our algorithms enable us to handle both large numbers of observations per second as well as large merged graphs. In particular, on a cluster with 9 compute nodes, PASS can reliably handle between 400K and 569K observations per second and merged graphs with as many as 50K vertices.

Overlapping community is a phenomenon often observed in numerous real-world networks. Fire-spread (Pattanayak et al. in Swarm Evol Comput. 44: 1–48 (2019) [11]) community detection algorithm is an efficient algorithm to detect overlapping community structures. In this work, the Fire-spread algorithm is modified to establish a relationship between community structure and clustering coefficient. By using different networks and executing the modified Fire-spread algorithm, it is found that the clustering coefficient is highly correlated with community structure. Finally, a simpler community detection algorithm, derived from the fire-spread algorithm, is proposed, where the clustering coefficient is used as a threshold value. To validate the proposed algorithm, it is compared with some state of art community detection algorithms based on the NMI score.

Community detection is one of the primary tools to discover useful information that is hidden in complex networks. Some community detection algorithms for bipartite networks have been proposed from various viewpoints. However, the performance of these algorithms deteriorates when the community structure becomes unclear. Enhancing community structure remains a nontrivial task. In this paper, we propose a community detection algorithm, called ECD, that enhances community structure in bipartite networks. In the proposed ECD, the topology of a network is modified by reducing unnecessary edges that are connected to neighboring low-weight communities. Therefore, an ambiguous community structure is converted into a structure that is much clearer than the original structure. The experimental results on both artificial and real-world networks verify the accuracy and reliability of our algorithm. Compared with existing community detection algorithms using state-of-the-art methods, our algorithm has better performance.

The research on the structural characteristics of Heterogeneous combat network (HCN) can provide an effective basis for its scientific organization and guarantee. The community structure analysis based on complex network can provide a powerful method for the research on the structural characteristics of HCN. In view of this, this paper put forward the idea of studying the structural characteristics of HCN based on community structure, and improve Hete_M community detection algorithm. In the experiment, different functional nodes are used as seeds to detect the community structure, which can provide reference for the node configuration of different functional elements. Meanwhile, the results show that the effect of dividing HCN structure with reconnaissance perception nodes as seeds is better, which provide a new idea for organizing HCN with reconnaissance perception elements as the core.

Previous studies unveiled that huge channel capacity can be achieved by employing multi-element antenna arrays at both transmitter and receiver. Foschini et.al. proposed a vertical Bell Laboratories layered space-time (V-BLAST) system with a detection algorithm based on decision feedback. This system is computationally efficient and yet supporting a high data rate. However, the detection algorithm shows unsatisfactory performance when an equal number of antennas are employed at both ends. In this paper, a parallel detection algorithm is proposed to improve the performance in such scenarios. The detector consists of many low complexity sub-detectors, all of them operating in parallel on a sub-channel matrix. By selecting the submatrix appropriately the overall performance can be improved substantially.

Networks have become a widely used tool for modeling complex systems in many different fields. This approach is extremely useful for representing interactions among genes, social relationships, Internet communications or correlations of prices within a stock market, to name just a few examples. By analyzing the structure of these networks and understanding how their different elements interact, we could improve our knowledge of the whole system. Usually, nodes that compose these networks tend to create tightly knit groups. This property, of high interest in many scientific fields, is called community structure and improving its detection and characterization is what this thesis is all about. The first objective of this work is the generation of efficient methods able to characterize the communities of a network and to understand its structure. Second, we will try to create a set of tests where such methods can be studied. Finally, we will suggest a statistical measure in order to be able to properly assess the quality of the community structure of a network. To accomplish these objectives, first, we generate a set of algorithms that can transform a network into a hierarchical tree and, from there, to determine their most relevant communities. Furthermore, we have developed a new type of benchmarks for effectively testing these and other community detection algorithms. Finally, and as the most important contribution of this work, it is shown that the community structure of a network can be accurately evaluated using a hypergeometric distribution-based index. Thus, the maximization of this measure, called Surprise, appears as the best proposed strategy for detecting the optimal partition into communities of a network. Surprise exhibits an excellent behavior in all networks analyzed, qualitatively outperforming any previous method. Thus, it appears as the best measure proposed to this end and the data suggests that it could be an optimal strategy to determine the quality of the community structure of complex networks.

Online social media network has become part of human life by transforming the way users create new social relations or relate with family and friends. Online social network (OSN) has drastically increased the rate at which people interact with each other by simplifying the means of communication. However, privacy is raising a serious concern. All user generated data within the OSN system need to be protected against unauthorized friends or hackers or even against the provider of OSN. Many research works are going on to encounter the privacy issues in OSN. This paper analyses the limitations of the recent work being done in this field and proposes an efficient abstract solution to them.

We examine the problem of effective identification of community structure of a network whose elements and their respective relationships manifest through streams. The problem has recently garnered much interest as it appears in emerging computational environments and concerns critical applications in diverse areas including social computing, web analysis, IoT and biology. Despite the already expended research efforts in detecting communities in networks, the unprecedented volume that real-world networks now reach, renders the task of revealing community structures extremely burdensome. The sheer size of such networks oftentimes makes their representation in main memory impossible. Thus, processing the developing graphs to extract the underlying communities remains an open challenge. In this paper, we propose a graph-stream community detection algorithm that expands seed-sets of nodes to communities. We consider a stream of edges and aim at processing them to form communities without maintaining the entire graph structure. Instead, we maintain very limited information regarding the nodes of the graph and the communities we seek. In addition to our novel streaming approach, we both develop a technique that increases the accuracy of our algorithm considerably and propose a new clustering algorithm that allows for automatically deriving the size of the communities we seek to detect. Our experimental evaluation using ground-truth communities for a wide range of large real-word networks shows that our proposed approach does achieve accuracy comparable or even better to that of state-of-the-art non-streaming community detection algorithms. More importantly, the attained improvements in both execution time and memory space requirements are remarkable.

The technology is making its road of development continuously and all the users are enjoying the benefits from it and leading an easy life. With the rise in usage of internet, many people join online social networks to connect with the world, share their views, and interact with others. Thus, the increasing sizes of social networks are becoming complex and are involving a huge amount of data. Mining these data has become one of the domains in information technology. The fundamental application and/or also active research from this data is discovering meaningful communities from interaction among different people of a huge network. The algorithms on community detection are very useful & necessary to find the pattern or structure of the social network, analyze the social network, find out the relationship between two people and also control the sentiments of the people. The quality of detected communities is also measured by using different parameters. This work starts with an explanation on the concepts of online social networks, communities and community detections then survey on the different community detection algorithms have been done based on the different categories.

In the real network, there must be a large and complex network. The solution to understand that kind of network structure is using the community detection algorithms. There are a lot of other algorithms out there to perform community detection. Each of the algorithms has its own advantages and disadvantages with different types and scale of complex network. The Louvain has been experimented that shows bad connected in community and disconnected when running the algorithm iteratively. In this paper, two algorithm based on agglomerative method (Louvain and Leiden) are introduced and reviewed. The concept and benefit are summarized in detail by comparison. Finally, the Leiden algorithm’s property is considered the latest and fastest algorithm than the Louvain algorithm. For the future, the comparison can help in choosing the best community detection algorithms even though these algorithms have different definitions of community.