Time-series models could be applied to capture the dynamic of network change from theoretical assumptions on lag numbers. I did a summary from one paper taking care of this problem. Check this out Modeling History Dependence in Network-Behavior Coevolution.
I am a PhD student in the School of Journalism and Mass Communication of UW-Madison. I like to apply advanced data mining techniques to answering theoretical research questions, particularly for interpreting machine learning algorithms as “common knowledge”. One major difference between algorithm learning and statistical modeling is the generalizability the two tools afford! As traditional statistics impose strict mathematical assumptions and we are either unaware of it or puzzled by assorted diagnostic tools, it is essential to let machine do the work and give us more time to focus on theoretical thinking, or living! Blogs about popular data science application, along with my research, will be posted. Here comes the first source I found intriguing: Ideas on Interpreting Machine Learning.