Report title: Data-driven selection of the number of change-points
Speaker: Professor Changliang Zou Nankai University
Reporting time: 13:30-14:20, January 11, 2021
Report location: Zoom meeting (Zoom meeting id: 770 311 8512, password: 378548)
School contact: Wang Peijie firstname.lastname@example.org
In multiple change-point analysis, one of the main difficulties is to determine the number of change-points. In this talk, I will introduce a data-driven approach based on an order-preserved sample-splitting strategy. Under a unified framework, a cross-validation estimation scheme is developed to achieve consistent selection. Furthermore, we construct a simple yet effective selection procedure which can quantify “uncertainty”, say controlling certain error rate. The proposed methods are applicable to most kinds of popular change -point algorithms.
Professor of the School of Statistics and Data Science, Nankai University. He received his Ph.D. degree from Nankai University in 2008 and then stayed on to teach. Mainly engaged in statistics and its cross-research and practical application in the field of data science. Research interests include: high-dimensional data statistical inference, large-scale data flow analysis, change point and abnormal point detection, etc., in Ann.Stat., Biometrika, J.Am.Stat.Asso., Math. Program., Technometrics, IISE Tran He has published dozens of papers in journals in the field of statistics and industrial engineering, and presided over major projects of the National Natural Science Foundation of China, outstanding youth projects, outstanding youth projects, etc.