A review of similarity measures and link prediction models in social networks

Abstract

Social network is a web-based platform which enables people to share information, make new connections and explore various events that occur in society. In social networks, link prediction techniques are widely used to discover new indirect relationships that may occur in the future. These techniques are also utilized to effectively detect missing links in any monitored network. This study presents a concise review of the similarity measures, techniques employed in predicting future links and application of link prediction with emphasis on dynamic networks. An analysis of available models for link prediction and their suitability for heterogeneous, large, static or dynamic networks is also presented.