CSG Cluster: A Collaborative Similarity Based Graph Clustering for Community Detection in Complex Networks
Smita S Agrawal1, Atul Patel2
1Prof Smita S Agrawal, Department of CSE, Institute of Technology, Nirma University, Ahmedabad (Gujrat), India.
2Dr Atul Patel, Dean of CMPICA, CHARUSAT University, Changa (Gujarat), India
Manuscript received on 18 June 2019 | Revised Manuscript received on 25 June 2019 | Manuscript published on 30 June 2019 | PP: 1682-1687 | Volume-8 Issue-5, June 2019 | Retrieval Number: E7475068519/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Many real-world social networks exist in the form of the complex network, which includes very large scale structured or unstructured data. The large scale networks like brain graph, protein structure, food web, transportation system, WorldWide Web, online social network are sparsely connected globally anddensely connected locally. For detectingdensely connected clusters from complex networks, graph clustering methods are useful. Graph clustering performs through partition a graph based on edge cut, vertex cut, edge betweeness, vertex similarities, topological structure of graph. Most of the graph clustering methods predominantly emphasis on topological structure of graph and not bearing in mind the vertex properties/attributes or similarity based onindirectly connected vertices. In this paper, we propose a CSGCluster, a novel collaborative similarity based graph clustering methodfor community detection ina complex network. In this, we introduce concepts, Approachable Unitto find similaritiesfor directly connected vertices andintroduced shortest path strategy for indirectlyconnected vertices and based on that a graph clustering method,CSG-Cluster is presented. For this, a new collaborative similarity approach is adopted to computevertex similarities. In the CSG-Cluster method, weform a group of vertices based on distance measures based on calculated similarity with the help of K-Medoids framework. Performs experiment on two real datasets with other relevant methods in whichresults shows the effectiveness of CSG-Cluster. This idea is suitable for graph database to apply collaborative similarity during query processing
Keywords: Attribute Similarity, Community Detection, Complex Network, Graph Clustering, Vertex Similarity
Scope of the Article: Computer Graphics