报告题目：Community Detection for Large-scale Networks

主讲人：王江洲

单位：深圳大学

时间：10月29日15:00

地点：学院二楼会议室

摘要：Statistical analysis of network science is a thriving research field with broad applications in biology, informatics, sociology, and computer science. Many complex systems from diverse domains can be represented as networks, which enhance our understanding of these systems. A fundamental task of network analysis is detecting the community structure where nodes cluster together. For this task, a great many of methods have been proposed. However, the existing methods often fall short in handling large-scale networks and ensuring accuracy, efficiency, convergence, and statistical consistency. To overcome these difficulties, we introduce two novel likelihood-based inference frameworks, profile pseudo-likelihood and split-likelihood, and the corresponding algorithms for large networks，and prove their theretical properties. Furthermore, for the multi-layer attributed networks (a type of network data with more complex structure), we introduce the penalized alternating factorization (PAF) algorithm detect their community structures.

简介：王江洲，深圳大学数学科学学院统计系助理教授。2021年博士毕业于东北师范大学，师从郭建华教授。随后于2021～2023年在南方科技大学统计与数据科学科学系做博士后，合作导师为荆炳义教授和邵启满教授。主要研究方向为大规模网络数据的统计分析和大规模相依数据的多重检验。目前在统计学领域期刊发表SCI论文7篇，其中包括：Journal of the American Statistical Association、Computational Statistics and Data Analysis、Computational Statistics和Stat等统计学国际期刊。主持国家自然科学青年基金项目、中国博士后科学基金面上项目和中国博士后科学基金特别资助（站中）项目各1项。同时也曾多次受邀在ICSA国际会议上做报告，并担任Statistica Sinica、Computational Statistics & Data Analysis和Statistics and Its Interface等多个统计学期刊的审稿人。