Network Structural Analysis Based on Graphical Measures and Metrics

Abstract

In the current era of rapid information exchange within a communication network like social media, computer networks, etc., identifying the communication pattern and densely interconnected region and analyzing it can yield noteworthy information. This information gathering may help in identification of suspicious group activities, identifying any anomaly in the group behavior, etc. In this paper, use of basic graph theory ideas and concepts has been made to identify densely interconnected region and patterns of communication within and between different communities. Further, this information will be utilized to identify any deviation from the general trend which may give us scope to further investigate the region of network. Such studies may be employed upon social media networks to identify suspicious groups. Based on general graph metrics analysis like motifs, community detection, average degree distribution, etc., various scenarios will be depicted to see how the evaluated metrics can be used to look for the solutions of the problem. The graphical models will be mimicked using SNAP (Simulation and Networks Analytics tool) and machine learning algorithm to evaluate the results depending upon various scenarios. Finally, the paper will conclude with the analysis of the results obtained on the aspect that how different situations can be handled based upon the network metrics.