时 间：2021年10月14日13:00-14:30
地 点：腾讯会议：509 200 457
题 目：Community Influence Analysis in Social Networks
报告人：兰伟 西南财经大学副教授
主持人：方方 教授
主 办：统计与数据科学前沿理论及应用教育部重点实验室
摘 要：
Heterogeneous influence detection across network nodes is an important task in network analysis. This paper proposes a community influence model (CIM) by assuming that the nodes can be classified into different communities (i.e., clusters or subgroups) and the nodes within the same community share the common influence parameters. Employing the quasi-maximum likelihood approach, together with the fused lasso-type penalty, we can not only identify the number of communities, but also estimate the influence parameters, without imposing any specific distribution assumption on the error terms. We further demonstrate the resulting estimators enjoy the oracle properties; namely, they perform as well as if the true underlying network structure were given in advance. The proposed approach is also applicable to identify influence nodes under homogeneous setting. To assess the adequacy of the homogeneous influence, the likelihood-ratio type test and its asymptotic theory are established. The performance of our methods is illustrated via simulation studies and two empirical examples on stock data and on coauthor citations for statistical journals.
报告人简介：
兰伟，博士毕业于北京大学光华管理学院，现为西南财经大学副教授，博士生导师。主要研究方向为高维数据建模、大型网络数据分析和投资组合优化。主持国家自然科学基金项目和多个重点重大项目子课题。在JASA,AOS, JOE, JBES，《金融研究》等国内国际知名期刊发表和接收中英文论文40余篇。