Identification of Influential Nodes in Social Network: Graph Approach

Abstract

The identification of nodes that are actually influencing other nodes is one of the core and fundamental research problems in the social network data sciences. There are multiple proven techniques available globally that depict the influence of the root node on the remaining nodes in the given network. Although methods and methodology is already proven and multiple experiments have been carried out in this area, the current research work is an extension of the influential node identification problem with the perspective of graph traversal. To prove the methodology and new approach a tree and AUC (area under the curve) - based analysis method has been proposed, TArank in which the information collected from the breadth-first search tree is used to identify influential nodes. From the obtained results and analysis, the effectiveness of TARank in the identification of influential nodes has been validated. The concentration of data at a particular segment of the plot is observed. Additionally, the tree view of the analysis predicts the root and leaf node thus helping in the identification of the most influential nodes in the online social media network. Keywords: Influential Node, PCC, TARank, Degree Centrality, PCC