Enhancing Sparse Adjacency Matrix for Community Detection in Large Networks

Authors

  • Ali Falah Yaqoob Department of Computer Science, College of Science, University of Baghdad, Baghdad-Iraq
  • Basad Al-Sarray Department of Computer Science, College of Science, University of Baghdad, Baghdad-Iraq

Keywords:

Community Detection, Tabu Search, Fuzzy c-mean (FCM), Laplacian graph

Abstract

This paper presents a problem of community detection in sparse network. Graph represents the network with {0, 1} symmetric matrix, this matrix is defined to be sparse when most of its entries are zeros. The problem of community detection of this type of networks is non-deterministic polynomial-time hardness (NP-hard) problem. Here, we give a simple idea to regularize the sparse matrix by adding a heuristic parameter to the entries of the matrix. This work performs integrating Tabu Search via Fuzzy C-mean to compute variants of the modularity maximization. The results show the ability of the proposed method to define structure of the network by optimizing different types of the quality functions; the results show the global function gives the high value in must runs when apply it on a large sparse real networks.

Published

2019-12-01

Issue

Section

Articles

How to Cite

(1)
Enhancing Sparse Adjacency Matrix for Community Detection in Large Networks. ANJS 2019, 22 (4), 75-85.