TY - JOUR
T1 - Community detection in complex networks using multi-objective bat algorithm
AU - Doush, Iyad Abu
AU - Alrashdan, We’am Bilal
AU - Al-Betar, Mohammed Azmi
AU - Awadallah, Mohammed A.
N1 - Publisher Copyright:
Copyright © 2020 Inderscience Enterprises Ltd.
PY - 2020
Y1 - 2020
N2 - Community detection is the problem of identifying communities in which we aim to discover groups of nodes with high connectivity within the same group and with low connectivity outside the group. Community detection is considered to be a non-deterministic polynomial-time hard problem. Heuristic algorithms can be used to solve such a complex optimisation problem. Bat algorithm (BA) is a meta-heuristic optimisation algorithm. The BA can be used to model a multi-objective optimisation problem. In this paper, the multi-objective bat algorithm (MOBA) is adapted to model and solve the community detection problem. In order to evaluate the algorithm, four real-world datasets are used. The performance of the algorithm is compared with seven other methods from the literature. The comparison was in terms of two metrics to check the quality of the obtained community namely modularity (Q) and normalised mutual information (NMI). The results show that the proposed algorithm outperforms all algorithms in one dataset and that it is competitive in other cases.
AB - Community detection is the problem of identifying communities in which we aim to discover groups of nodes with high connectivity within the same group and with low connectivity outside the group. Community detection is considered to be a non-deterministic polynomial-time hard problem. Heuristic algorithms can be used to solve such a complex optimisation problem. Bat algorithm (BA) is a meta-heuristic optimisation algorithm. The BA can be used to model a multi-objective optimisation problem. In this paper, the multi-objective bat algorithm (MOBA) is adapted to model and solve the community detection problem. In order to evaluate the algorithm, four real-world datasets are used. The performance of the algorithm is compared with seven other methods from the literature. The comparison was in terms of two metrics to check the quality of the obtained community namely modularity (Q) and normalised mutual information (NMI). The results show that the proposed algorithm outperforms all algorithms in one dataset and that it is competitive in other cases.
KW - Bat algorithm
KW - Community detection
KW - MOBA
KW - Multi-objective bat algorithm
KW - Multi-objective optimisation
UR - http://www.scopus.com/inward/record.url?scp=85083384098&partnerID=8YFLogxK
U2 - 10.1504/IJMMNO.2020.106529
DO - 10.1504/IJMMNO.2020.106529
M3 - Article
SN - 2040-3607
VL - 10
SP - 123
EP - 140
JO - International Journal of Mathematical Modelling and Numerical Optimisation
JF - International Journal of Mathematical Modelling and Numerical Optimisation
IS - 2
ER -