Community Detection Based on Local Information and Dynamic Expansion

Luo, Yongping; Wang, Li; Sun, Shiwen; Xia, Chengyi

doi:10.1109/access.2018.2878674

Cited by 17 publications

(5 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Luo et al [16] proposed two signed modularity functions to evaluate the community properties in networks. Using the proposed modularity functions, they proposed a dynamic expansion algorithm based on local information for global community detection.…”

Section: Related Workmentioning

confidence: 99%

Large-Scale Network Community Detection Using Similarity-Guided Merge and Refinement

Tunalı

2021

IEEE Access

View full text Add to dashboard Cite

It is possible to extract valuable insights about the functional properties of a system by identifying and inspecting the community structure in the network that models the system. Community detection aims to extract these community structures from networks. Many community detection methods have been proposed that consider the problem from different perspectives. However, with the emergence of very large and complex networks from variety of domains, there has been a growing need for community detection methods that can operate at scale effectively and efficiently. Considering this, we propose a novel algorithm for large-scale community detection, based on two novel similarity indices we propose as well.In the first stage of our proposed algorithm, we generate candidate communities using a mechanism similar to information propagation very rapidly. Then, we merge small candidates that have fewer nodes than a calculated threshold with the larger ones using similarity between nodes and communities. Next, we engage a refinement operation on the candidates by moving all nodes to the candidates to which they are most similar using the same similarity index again. After that, we merge small communities with larger ones by using the similarity between communities until no gain in the modularity is obtained. Finally, in the last stage, we employ the same refinement operation as in the third stage. With an extensive experimentation on real-world and artificially-generated benchmark networks, we demonstrate and verify the performance and effectiveness of the proposed algorithm comparing it with the state-of-the-art methods. Experimental results indicate that our algorithm scales very well with growing size and complexity of networks. Besides, our algorithm outperforms most state-of-the-art community detection methods both in detection performance and computation time.

show abstract

Section: Related Workmentioning

confidence: 99%

Large-Scale Network Community Detection Using Similarity-Guided Merge and Refinement

Tunalı

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…However, some local methods may lack access to global knowledge, resulting in unfavorable accuracy. Local algorithms typically have a time complexity that is close to linear and are considered to have an acceptable time complexity for large-scale network analysis [5].…”

Section: Introductionmentioning

confidence: 99%

CDBRA: Community Detection Based on Random Algorithm in Social Networks

Suganthini,

Baskaran,

Dhivya

2023

JECNAM

View full text Add to dashboard Cite

Understanding the topology and functions of complex networks allows us to derive valuable information from them. There are various types of these networks. Community detection is a significant research area that involves dividing a network graph into subsets of nodes, known as communities. Each community consists of nodes that have dense communication with each other and sparse communication with nodes outside the community. This work proposes the use of Community Detection based on random Algorithm (CDBRA) to identify novel communities with low complexity and high accuracy by using both local and global network information. The proposed method consists of four components: Pre-Processing, Node Identification, Intra-Community Structure, and Inter-Community Structure. In the initial component, the task involves recognizing and saving similarity measures. Additionally, it requires assigning suitable weights to network vertex and edges, taking into the account of local and global network information. The next level involves using a random algorithm enhanced by nodes' weights to determine similarity measures for Node Identification. The third level, Intra-Community Structure, aims to achieve various community structures. The fourth level ultimately chooses the optimal community structure by taking into account the Inter-Community Structure and the evaluation functions derived from network’s local and global information. To assess the proposed method on various scenarios involving real and artificial networks. The proposed method outperforms existing methods in detecting community structures similar to real communities and provides efficient evaluation functions for all types and sizes of networks.

show abstract

“…Recently, a few excellent studies have been published on community detection based on local information and dynamic expansion [3]; the application of random matrix theories, and graphs or networks [4,5]. Because each method has its strong points and weaknesses, we propose to combine the strong points and reduce the limitations or the weak points and use a combination of these methods for finding the correlations between agents in the stock market system.…”

Section: Introductionmentioning

confidence: 99%

Community Detection of Dynamic Complex Networks in Stock Markets Using Hybrid Methods (RMT‐CN‐LPAm+ and RMT‐BDM‐SA)

Purqon

Jamaludin

2021

Front. Phys.

View full text Add to dashboard Cite

A stock market represents a large number of interacting elements, leading to complex hidden interactions. It is very challenging to find a useful method to detect the detailed dynamical complex networks involved in the interactions. For this reason, we propose two hybrid methods called RMT-CN-LPAm+ and RMT-BDM-SA (RMT, random matrix theory; CN, complex network; LPAm+, advanced label propagation algorithm; BDM, block diagonal matrix; SA, simulated annealing). In this study, we investigated group mapping in the S&P 500 stock market using these two hybrid methods. Our results showed the good performance of the proposed methods, with both the methods demonstrating their own benefits and strong points. For example, RMT-CN-LPAm+ successfully identified six groups comprising 485 involved nodes and 17 isolated nodes, with a maximum modularity of 0.62 (identified more groups and displayed more maximum modularity). Meanwhile, RMT-BDM-SA provided useful detailed information through the decomposition of matrix C into Cm (market-wide), Cg (group), and Cr (noise). Both hybrid methods successfully performed very detailed community detection of dynamic complex networks in the stock market.

show abstract

Community Detection Based on Local Information and Dynamic Expansion

Cited by 17 publications

References 61 publications

Large-Scale Network Community Detection Using Similarity-Guided Merge and Refinement

Large-Scale Network Community Detection Using Similarity-Guided Merge and Refinement

CDBRA: Community Detection Based on Random Algorithm in Social Networks

Community Detection of Dynamic Complex Networks in Stock Markets Using Hybrid Methods (RMT‐CN‐LPAm+ and RMT‐BDM‐SA)

Contact Info

Product

Resources

About