Authors: Akula, Ramya, Yousefi Niloofar, and Garibay Ivan
Abstract
Information spreads on complex social networks extremely fast. A piece of information can go viral in no time and can be harmful. Often it is hard to stop this information spread causing social unrest. An intentional spread of software vulnerabilities in GitHub has caused millions of dollars in losses. GitHub is a social coding platform that enables a huge number of open source software projects to thrive. To better understand how the information spreads on GitHub, we develop a deep neural network model: “DeepFork”, a supervised machine learning based approach that aims to predict information diffusion; considering the node as well as topological features in complex social networks. In our empirical studies, we observed that information diffusion can be detected by link prediction using supervised learning. This model investigates the followee-follower influence that underlay information dynamics in social coding platform. DeepFork outperforms other machine learning models as it better learns the discriminative patterns from the input features. DeepFork helps us in understand the human influence on information spread and evolution.