统计与数据科学系系列学术报告之四百八十三期

时间： 2025年9月8日（周一）16:00-17:00

主持人：复旦大学管理学院统计与数据科学系蒋斐宇副教授

地点：史带楼302室

报告人：Dr. Fan Wang 王凡

The University of Melbourne

题目：Change Point Analysis in Dynamic Multilayer Networks

摘要：In this talk, I will introduce the multilayer random dot product graph (MRDPG) model, a generalization of the random dot product graph to multilayer networks. A tensor-based methodology is developed for estimating edge probabilities and is shown to outperform existing approaches. In the dynamic setting, an online change point detection framework is formulated, where each time point provides a realization from an MRDPG. A novel nonparametric algorithm, combining density kernel estimators with tensor-based techniques, is proposed. The method applies broadly to network models, including stochastic block models as special cases. From a theoretical perspective, the procedures minimize detection delay while controlling false alarms. Extensive numerical experiments, including an application to U.S. air transportation networks, corroborate the theoretical results.

个人简介：Fan Wang is a Research Fellow at the Centre for Research in Statistical Methodology (CRiSM) in the Department of Statistics at the University of Warwick. Her research focuses on change point analysis, statistical network analysis and transfer learning. She will be joining the School of Mathematics and Statistics at the University of Melbourne as a Lecturer at the end of this year. Prior to her postdoctoral position, she obtained her PhD in Statistics from the University of Warwick.