报告时间：2018年10月19日 14:00-15:30
报告地点：明德主楼1016会议室
报告主题：Consistent estimation of the number of change-points
报告摘要：In multiple change-point analysis, one of the major challenges is to estimate the number of change-points. In this talk, I will briefly introduce some recent developments on data-driven selection criteria from two main aspects, parametric and semi-parametric modelling. Under a unified parametric framework, an order-preserved sample-splitting strategy is shown to achieve asymptotic selection consistency under some mild conditions. For a broad class of semi-parametric regression models, we propose a joint least-squares loss function which serves as a reasonable measure for determining the number of change-points. Some extensions such as the high-dimensional settings, are also involved.
个人简介：邹长亮，南开大学统计与数据科学学院教授、博导。08年于南开大学获博士学位，随后留校任教。研究方向是统计学、质量科学及其与数据科学领域的交叉研究和实际应用。研究兴趣包括：高维数据统计推断、大规模数据分析、统计过程控制、变点和异常点检测等，在统计学和质量科学领域杂志上发表论文几十篇。