A generic topology discovery approach for huge social networks

Abstract

Social Networks are gaining importance due to their enablement of modeling various types of interactions among individuals, communities and organizations. Network Topologies play a major role in analyzing the social networks for a variety of business application scenarios such as finding influencers in product campaigning and virtual communities to recommend music downloads. Social networks are dynamic in nature and detection of topologies from these networks presents a host of new challenges. In this paper, we present approaches for topology discovery, particularly star, ring and mesh, based on the measures of network centrality. These approaches facilitate an efficient way of discovering topologies for analyzing large social networks. We also discuss experiments on DBLP dataset to show the viability of our proposed approach.

Big Data Search and Mining

“…In the next section, we discuss clustering social networks using topology discovery (e.g., star, ring, and mesh topologies) [19]. The computed topology scores are used to determine the extent to which the clusters grow [4].…”

“…Discovering network topologies in social networks would provide tremendous benefits for business applications such as finding key influencers for product campaigning and identifying virtual communities to recommend music downloads [4]. For example, star topology is useful to find key influencers who can motivate a large number of people; ring topology would help in determining critical nodes for data transfer; mesh topology would help in categorizing active communities/fully connected components which are candidates for promoting new offers [19].…”

“…Social network measures such as betweenness, closeness, and degree centrality, clustering coefficient and eccentricity address various types of roles/connections that exist among vertices [19]. These measures in social network analysis give a rough indication of the social power of a vertex based on the pattern in which the vertices of a network are connected and extensively used directly/indirectly in developing most of the social network applications.…”

“…Therefore, the lesser the eccentricity, the more the influencing power of a vertex. The above five measures are considered in [19] to discover star, ring, and mesh topologies. Note that there also exist several other network centrality measures (e.g., radius centrality, eigenvalue centrality, and alpha centrality).…”

A Scalable Algorithm for Discovering Topologies in Social Networks

“…They learnt the two parameters neighborhood distance and min-points which lead to better cluster quality. Krishna et al [13] presented an approach to discover community structures in terms of network topologies using the centrality measures such as degree and betweeness centrality. Hajeer et al [9] focused on creating social communities using Genetic algorithms, where the fitness function is based on the intra cluster quality.…”