Extended Feature-Driven Graph Model for Social Media Networks

Abstract

One of the major problems in social media research on networks is the lack of empirical datasets. To avoid this problem, different graph generation models have been proposed to represent real social graphs to an acceptable extent. This enables researchers to try and evaluate new methods on a large number of social media networks. The work described here aims to introduce an extended feature-driven model that provides synthetic graphs that are sufficiently representative of real retweet-based graphs generated from Twitter datasets. We have used three real retweet-based graphs introduced from three Twitter datasets, and structure-based graph metrics (statistical similarity metrics) to evaluate the performance of our extended model. Our experimental results demonstrate that our extended model stands out as a useful model to provide a highly accurate representation of the real-world graphs over other proposed feature-driven models.