Jakšić Kruger, Tatjana
Mathematical Institute of the Serbian Academy of Sciences and Arts
|Title:||Variable Neighborhood Search Approach to Community Detection Problem||Series/Report no.:||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)||Volume:||LNCS 13858||First page:||188||Last page:||199||Conference:||International Conference NMA 2022: Numerical Methods and Applications||Issue Date:||1-Jan-2023||Rank:||M33||ISBN:||978-3-031-32411-6||ISSN:||0302-9743||DOI:||10.1007/978-3-031-32412-3_17||Abstract:||
Community detection on graphs can help people gain insight into the network’s structural organization, and grasp the relationships between network nodes for various types of networks, such as transportation networks, biological networks, electric power networks, social networks, blockchain, etc. The community in the network refers to the subset of nodes that have greater similarity, i.e. have relatively close internal connections. They should also have obvious differences with members from different communities, i.e. relatively sparse external connections. Solving the community detection problem is one of long standing and challenging optimization tasks usually treated by metaheuristic methods. Thus, we address it by basic variable neighborhood search (BVNS) approach using modularity as the score for measuring quality of solutions. The conducted experimental evaluation on well-known benchmark examples revealed the best combination of BVNS parameters. Preliminary results of applying BVNS with thus obtained parameters are competitive in comparison to the state-of-the-art methods from the literature.
|Keywords:||Metaheuristics | Modularity maximization | Optimization on graphs | Social networks||Publisher:||Springer Link|
Show full item record
checked on Dec 7, 2023
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.