Article Preview
TopIntroduction
The rapidly rate of our life today needs fast interactions and response. Social networks (SN) used as fastest way for communication purpose as we can exchange information, files, voice and videos. The simplest definition of social networks is set of nodes and edges, where nodes represent any type of objects and edges are the relations between it (Rossi et al., 2015; Xu et al., 2016). Individuals in social networks may be have tendency to create a group to share specific topic, interest or other goal. 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 detection in social networks and graph analysis are important researches in recent years as its need in many field as marketing, energy saving and diseases (Wang et al., 2015). Community Detection 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.
By using community structure in networks analysis, several researches find that many areas are affected by community structure such as bio-informatics, sociology and information science (Babers et al., 2015). In Network field, clustering web clients who have similar interests and are near to each other geographically. The performance of services which provided to them can be improved by dedicating a mirror server for them (Krishnamurthy et al., 2000) In marketing field, identifying group of customers which have similar tendency toward purchasing specific products enables to set up recommendation list to guide them through the marketing and therefore enhance the business opportunities and increase revenue (Reddy et al., 2002).
Metaheuristic algorithms inspired by nature are became popular in offering new good solutions for some problems. Machine learning uses optimization algorithms to enhance the performance of learning and optimization get its thinking from machine learning. The searching for optimal solutions to a particular problem is defined as optimization process and it uses several agents (Kumar et al., 2016; Yang & Deb, 2014; Dey et al., 2014; Dey & Ashour, 2016a). Metaheuristic algorithms characterized by its ability to imitate the best feature in nature by searching around the current best solution and select the best solutions and it also can explore the search space efficiently. Metaheuristic algorithms inspired by nature are used in this research to optimize the number of communities existing in datasets.
Cuckoo Search Optimization Algorithm, Lion Optimization Algorithm and Ant-Lion Optimization Algorithm presented in this research as metaheuristic inspired algorithms from nature for community detection in social networks. Comparative analysis between them was presented. In this research Zachary karate Club, The Bottlenose Dolphin Network, American College Football Network, and Facebook used as benchmark datasets for comparison to verify the efficiency of them.
The rest of this paper is organized as: Section 2 presents related work. Section 3 states methodology of the community detection problem and measures used for social network analysis. Section 4 presents the nature-inspired metaheuristic algorithms used in this research. Section 5 states the experimental results and a comparative analysis between the algorithms presented in this article. The conclusion of this research is states in section 6.