摘要:In this talk, we will introduce two self-normalized statistics to jointly test and sequentially detect mean and variance change-points in the long-memory time series. The limiting distributions of test statistics under the no change-point null hypothesis and three different alternative hypotheses are proved. In particular, a sieve bootstrap approximation method is proposed to determine the critical values. Extensive simulations indicate that the proposed tests perform well in finite samples, and can discriminate between mean and variance change-point. Finally, we illustrate our tests via three real data sets.

个人简介：陈占寿，青海师范大学数学与统计学院副院长，教授，博士生导师，青海省统计学会副会长，中组部“西部之光”访问学者，加拿大英属哥伦比亚大学访问学者，青海省高校“135高层次人才培养工程”拔尖学科带头人，青海省自然科学与工程技术学科带头人，青海省昆仑英才教学名师；主要从事时间序列变点分析，小区域估计等方面的研究工作，先后主持国家自然科学基金3项，青海省自然科学基金5项；发表科研论文50余篇，出版学术专著一部，获青海省自然科学优秀论文三等奖2项。