Kevin Lee
Department of Statistics
The Pennsylvania State University
Abstract:
Dynamic network modeling provides an emerging statistical technique for various real-world
applications. It is a fundamental research question to detect the community structure in dynamic networks. However, due to signifcant computational challenges and divculties in modeling communities, there is little progress in the current literature to effectively find communities
in dynamic networks. This talk introduces a novel model-based clustering framework for
dy namic networks, which is based on (semiparametric) exponential-family random graph models
and inherits the philosophy of finite mixture models. To determine an appropriate number of
communities, a conditional likelihood Bayesian information criterion is proposed. Moreover, an
efficient variational expectation-maximization algorithm is designed to solve approximate maximum
likelihood estimates of network parameters and mixing proportions. By using variational
methods and minorization-maximization techniques, our methods have appealing scalability for
large scale dynamic networks. Finally, the power of our methods is demonstrated in real applications
to dynamic international trade networks and dynamic arm trade networks. This talk is
based on joint works with Amal Agarwal, David Hunter and Lingzhou Xue.
Bio:
Kevin Lee is completing a Ph.D. in Statistics from Penn State University.
He earned a master's in statistics from Korea University in 2012.
His research interests include statistical analysis of network data, high-dimensional
statistical inference, and statistical machine learning and data mining.
---
All statistics graduate students are expected to attend.