DETECTION OF COHESIVE SUBGROUPS IN SOCIAL NETWORKS USING INVASIVE WEED OPTIMIZATION ALGORITHM

Abstract

Social network analysis (SNA) is a very popular research area that helps to analyze social structures through graph theory. Objects in social structures are represented by nodes and are modeled according to the relations (edges) they establish with each other. The determination of community structures on social networks is very important in terms of computer science. In this study, the Invasive Weed Optimization (IWO) algorithm is proposed for the detection of meaningful communities from social networks. This algorithm is proposed for the first time in community detection (CD). In addition, since the algorithm works in continuous space, it is made suitable for solving the CD problems by being discretized. The experimental studies are conducted on human-social networks such as Dutch College, Highland Tribes, Jazz Musicians and Physicians. The results obtained from experimental results are compared and analyzed in detail with the results of the Bat Algorithm and Gravitational Search Algorithm. The comparative results indicate that IWO algorithm is an alternative technique in solving CD problem in terms of solution quality.