Distinguishing between Different Mechanisms of Network Evolution using Network Motifs and Machine Learning

In this thesis, I will introduce two methodological tools for understanding the evolution of social networks. Using a mathematical representation of social networks based on covariance matrices, I demonstrate machine learning algorithms can use these representations of networks to properly classify networks based on their evolution mechanism with high accuracy. I also show the over-expression of particular network motifs can be used to distinguish between different network evolution processes.