Detecting community structure using label propagation with consensus weight in complex network

Liang, Zong-Wen; Li, Minglu; Yang, Fan; Petropulu, Athina P.

doi:10.1088/1674-1056/23/9/098902

Cited by 23 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The strategy to define such consensus partitions relies on observing whether a pair of vertices is in the same group in most of the partitions in the set. Studies (Lancichinetti & Fortunato, 2012), (Liang et al, 2014) and (Santos et al, 2016) obtained good results using these methods.…”

Section: Consensus Clusteringmentioning

confidence: 88%

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Tautenhain

Nascimento

2020

Expert Systems with Applications

View full text Add to dashboard Cite

Graph clustering is a challenging pattern recognition problem whose goal is to identify vertex partitions with high intra-group connectivity. This paper investigates a bi-objective problem that maximizes the number of intra-cluster edges of a graph and minimizes the expected number of inter-cluster edges in a random graph with the same degree sequence as the original one. The difference between the two investigated objectives is the definition of the well-known measure of graph clustering quality: the modularity. We introduce a spectral decomposition hybridized with an evolutionary heuristic, called MOSpecG, to approach this bi-objective problem and an ensemble strategy to consolidate the solutions found by MOSpecG into a final robust partition. The results of computational experiments with real and artificial LFR networks demonstrated a significant improvement in the results and performance of the introduced method in regard to another bi-objective algorithm found in the literature. The crossover operator based on the geometric interpretation of the modularity maximization problem to match the communities of a pair of individuals was of utmost importance for the good performance of MOSpecG. Hybridizing spectral graph theory and intelligent systems allowed us to define significantly high-quality community structures.

show abstract

Section: Consensus Clusteringmentioning

confidence: 88%

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Tautenhain

Nascimento

2020

Expert Systems with Applications

View full text Add to dashboard Cite

show abstract

“…The reason is mainly that of its use which does not require any parameters, easy implementation, fast execution for large networks, and the ability to detect valid communities in random graphs [15]. For that reason, studies using the algorithms continue, as conducted by [16]- [20].…”

Section: Label Propagationmentioning

confidence: 99%

Performance Analysis of Molecular Complex Detection in Social Network Datasets

Sulistianingsih¹,

Winarko²

2017

IJCA

View full text Add to dashboard Cite

Researches related graph dataset conducted for years. One of its main topics was community detection. The development of algorithms to do community detection continuously conducted by adjusting characteristics of datasets used. One of which is Molecular Complex Detection (MCODE) algorithm used to community detection in a dataset of protein-protein interaction (PPI). However, use of the algorithm still limited to PPI dataset only. The aim of this research was to conducted experiment usage of MCODE algorithm in other datasets such us social network datasets. An experiment conducted by comparing the performance of MCODE with other benchmark algorithms such us Label Propagation and Girvan-Newman. From the experiment performed was resulted that for modularity MCODE showed the best result when compared with others, followed Girvan-Newman and Label Propagation with its values were 0.67, 0.66 and 0.46, respectively. Furthermore, for a testing parameter such us running time and average clustering coefficient, MCODE showed better result compared with Girvan-Newman and Label Propagation. For running time, MCODE needed mean time as 0.053 s, GirvanNewman as 0.056 s and Label Propagation as 0.078 s and for test parameter of average clustering coefficient, MCODE was 0.37, Girvan-Newman was 0.44 and Label Propagation was 0.46.

show abstract

“…Assuming that a vertex is able to hold more than one label, LPA can be extended for overlapping community detection. There are a number of improved algorithms based on the LPA approach for overlapping community detection, such as COPRA(Community Overlap PRopagation Algorithm) [9], MLPA(Multi-Label Propagation Algorithm) [10], BMLPA(Balanced MLPA) [11], SLPA(SpeakerListener LPA) [12], LPAcw(LPA with consensus weight) [13], DLPA(Dominant LPA) [14], etc. They are differentiated by the way the selecting labels and their propagating strategies.…”

Section: A Lpamentioning

confidence: 99%

“…The strength is measured as the clustering coefficient of the vertex's neighbors belonging to the candidate community. We propose a new type of connection strength, the connection 4: for (each community C in local community structure LCS of the sub-network) do 5: for (each recently joining vertex JV of C) do 6: for (each external neighbor EN of JV) do 7: if (EN is not a member of C) then 8: SendMsg(EN,[operation=notifyCom,C]); 9: end if 10: end for 11: end for 12: end for 13: Fig. 3.…”

Section: ) Connection Strengthmentioning

confidence: 99%

“…In subsequent even super-steps, the algorithm first processes received community notification messages, then evolves communities of each vertex in a sub-network and finally is sending vertex actions on each local community to its external members to keep community replicas consistent. In subsequent odd super-steps, after processing received vertex action messages, community replicas in different sub-networks 1: 2: ALGORITHM: NotifyVertexCommunities(sub-network) 3: 4: for (each vertex V in the sub-network) do 5: for (each neighbor N of V) do 6: if (N is a local vertex) then 7: get local vertex N; 8: directly set V's communities for N; 9: else 10: SendMsg(N,[operation=notifyCommunity, V and its communities]); 11: end if 12: end for 13 NotifyVertexNeighbors(sub-network); 6: else if (super-step == 1) then 7: ProcessNeighborMessages(sub-network); 8: InitializeCommunities(sub-network); 9: NotifyNewCommunities(sub-network); 10: NotifyVertexCommunities(sub-network); 11: else if ((super-step % 2) == 0) and (super-step ≤ MAXSTEP)) then 12: ProcessNewCommunityMessages(sub-network); 13: ProcessVertexCommunityMessages(sub-network); 14: EvolveCommunities(sub-network); 15: NotifyVertexActions(sub-network); 16: else if ((super-step % 2) == 1) and (super-step ≤ MAXSTEP)) then 17: ProcessActionMessages(sub-network); 18: MergeCommunities(sub-network); 19: NotifyNewCommunities(sub-network); 20: NotifyVertexCommunities(sub-network); 21: else 22: ProcessActionMessages(sub-network); 23: Set all local vertexes to be halt; 24: end if 25: Fig. 8.…”

Section: E Alogrithm Frameworkmentioning

confidence: 99%

See 1 more Smart Citation

A Parallel Self-Organizing Community Detection Algorithm Based on Swarm Intelligence for Large Scale Complex Networks

Sun¹,

Jie²,

Sauer³

et al. 2017

2017 IEEE 41st Annual Computer Software and Applications Conference (COMPSAC)

View full text Add to dashboard Cite

Abstract-Community detection is a critical task for complex network analysis. It helps us to understand the properties of the system that a complex network represents and has significance to a wide range of applications. Nowadays, the challenges faced by community detection algorithms include overlapping community structure detection, large scale network analysis, dynamic changing of analyzed network topology and many more. In this paper a self-organizing community detection algorithm, based on the idea of swarm intelligence, was proposed and its parallel algorithm was designed on Giraph++ which is a semiasynchronous parallel graph computation framework running on distributed environment. In the algorithm, a network of large size is firstly divided into a number of small sub-networks. Then, each sub-network is modeled as a self-evolving swarm intelligence sub-system, while each vertex within the sub-network acts iteratively to join into or leave from communities based on a set of predefined vertex action rules. Meanwhile, the local communities of a sub-network are sent to other sub-networks to make their members have a chance to join into, therefore connecting these self-evolving swarm intelligence sub-systems together as a whole, large and evolving, system. The vertex actions during evolution of a sub-network are sent as well to keep multiple community replicas being consistent. Thus network communication efficiency has a great impact on the algorithm' performance. While there is no vertex changing in its belonging communities anymore, an optimal community structure of the whole network will have emerged as a result. In the algorithm it is natural that a vertex can join into multiple communities simultaneously, thus can be used for overlapping community detection. The algorithm deals with vertex and edge adding or deleting in the same way as the algorithm running, therefore inherently supports dynamic network analysis. The algorithm can be used for the analysis of large scale networks with its parallel version running on distributed environment. A variety of experiments conducted on synthesized networks have shown that the proposed algorithm can effectively detect community structures and its performance is much better than certain popular community detection algorithms.

show abstract

Detecting community structure using label propagation with consensus weight in complex network

Cited by 23 publications

References 27 publications

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Performance Analysis of Molecular Complex Detection in Social Network Datasets

A Parallel Self-Organizing Community Detection Algorithm Based on Swarm Intelligence for Large Scale Complex Networks

Contact Info

Product

Resources

About