Speaker: Shizhe Chen, Assistant Professor, Statistics, UC Davis
Title: "Semiparametric estimation for dynamic networks with shifted connecting intensities"
Abstract: To model time-evolving networks, researchers have extended the well-known stochastic block models to the dynamic setting, where the Bernoulli random variable that represents an edge becomes a point process and the static connecting probability between two nodes is generalized to the connecting intensity over time. However, in many applications, the connecting intensities are subject to node-wise time shifts that may cause unidentifiability or misclustering if not accounted for. In this talk, we propose a stochastic block model that incorporates the unknown time shifts in dynamic networks. We establish the conditions that guarantee the identifiability of cluster memberships of nodes and representative connecting intensities across clusters. Using shape invariant models, we propose computationally efficient semiparametric estimation procedures to simultaneously estimate time shifts, cluster memberships, and connecting intensities. We illustrate the performance of the proposed procedures via simulation experiments, and we further apply the proposed method on a neural data set to reveal distinct roles of neurons during motor circuit maturation in zebrafish.
Faculty web page: http://www.chenshizhe.com/
Seminar Time/Location: Thursday Sept 29, 4:10pm, at Mathematical Sciences Building 1147
Refreshments: 3:30pm, Courtyard of MSB 1147