A Nature-Inspired Metaheuristic Cuckoo Search Algorithm for Community Detection in Social Networks

A Nature-Inspired Metaheuristic Cuckoo Search Algorithm for Community Detection in Social Networks

Ramadan Babers, Aboul Ella Hassanien
DOI: 10.4018/IJSSMET.2017010104
(Individual Articles)
No Current Special Offers


In last few years many approaches have been proposed to detect communities in social networks using diverse ways. Community detection is one of the important researches in social networks and graph analysis. This paper presents a cuckoo search optimization algorithm with Lévy flight for community detection in social networks. Experimental on well-known benchmark data sets demonstrates that the proposed algorithm can define the structure and detect communities of complex networks with high accuracy and quality. In addition, the proposed algorithm is compared with some swarms algorithms including discrete bat algorithm, artificial fish swarm, discrete Krill Herd, ant lion algorithm and lion optimization algorithm and the results show that the proposed algorithm is competitive with these algorithms.
Article Preview

1. Introduction

Today, the fast way to contact person or organization became over the Internet by using social networks applications, social interactions over such applications became basic part of our daily activities. Social networks defined as set of nodes and edges, where nodes represent individuals or objects and edges represent the relationships or interactions between nodes (Xu et al. 2016; Rossi et al. 2015). Community is a term used to identify the tendency group of individuals to form condensed interactions and relations between them than individuals not belonging to that group (Bansal et al. 2011; Alzahrani & Horadam, 2016). Community structure used in networks analysis to achieve a specific target (Barabâsi et al. 2002), discover complex networks topology or discover valuable insights from the networks (Hassan et al. 2015). It used to detect groups and subgroups which have similar features (Xu et al. 2015; Kawamoto & Rosvall, 2015). There are several areas affected by community structure such as bio-informatics, sociology and information science (Babers et al. 2015). Several researches analysis complex networks and discover insights in it (Banati & Arora, 2015; Cao et al. 2015; Yu et al. 2015; Azar & Vaidyanathan, 2015; Emary et al. 2014a).

Traditional methods such as graph partitioning used to detect community within networks by dividing it based on predefined size. Spectral clustering method is based on similarity matrix and merging similar groups used by hierarchical clustering method (Choudhury & Paul). Community detection (CD) in social networks is one of the important researches in social networks and graph analysis (Wang et al. 2015). CD based on node centric which nodes in such community satisfy certain properties as degree and reachability, also each group in community satisfy certain condition as density of interaction within nodes in the same group (Azar & Vaidyanathan, 2015; Hassanien et al. 2014).

The network defined as group of objects or entities interact with each other and linked by relationships called edges. Nodes represent anything as person, cell phone, PC’s IP, car and product, the relationships between nodes are the connection linking them. There are different types of connections such that direct and indirect relation. The powerful tool for studying the structure of complex systems is network analysis (Albert & Barabási, 2002; Newman, 2003) which is useful in understanding the interactions and relationships established among nodes (Watts et al. 2002; Lusseau & Newman, 2004; Flack et al. 2006; Hassanien et al. 2015), also in describing and predicting the behavior of nodes.

Social Networks consist of two basics parts which are individuals (nodes) and relationships (edges) and can be defined formally using graph structure, IJSSMET.2017010104.m01 where IJSSMET.2017010104.m02 is the set of nodes and relations or edges are IJSSMET.2017010104.m03 (Hassan et al. 2015). The number of nodes denoted by IJSSMET.2017010104.m04 and the number of edges by IJSSMET.2017010104.m05. Given a set of nodes IJSSMET.2017010104.m06, the target is dividing those nodes into groups such that nodes in the same group are related and similar to each other and nodes in different group are dissimilar to each other. The challenge is to detect the number of communities within such complex networks, where each community is fulfilled objective quality function.

(1) Where IJSSMET.2017010104.m08 is an objective quality function and IJSSMET.2017010104.m09 is assuming that the quality measure minimized without lose in the context. Measures used for social networks analysis are different and depend on the network size. For example, the measures used: centrality, association index and node degree (Salama et al. 2012; Maharani & Gozali, 2015).

Complete Article List

Search this Journal:
Volume 15: 1 Issue (2024)
Volume 14: 1 Issue (2023)
Volume 13: 6 Issues (2022): 2 Released, 4 Forthcoming
Volume 12: 6 Issues (2021)
Volume 11: 4 Issues (2020)
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing