讲座题目：Hypothesis testing in high-dimensional censored-transformation models
内容摘要：With the rapid development of modern technologies, high-dimensional statistical inference of survival times has become increasingly important in various fields, including biostatistics and financial risk. Herein, we propose an efficient rank-based test statistic that is asymptotically normally distributed. The proposed test statistic allows for covariance-dependent censoring and is robust against heavy-tailed distributions and potential outliers. Furthermore, for practical purposes, we propose a new rank-based test statistic to test for the existence of high-dimensional features with high-dimensional control factors. Asymptotic distributions under the null hypothesis and local alternatives are established for the proposed test statistic. Numerical studies are performed to evaluate the finite-sample performance of the proposed test statistics. We illustrate the proposed method to an empirical analysis of a skin cutaneous melanoma (SKCM) dataset.
主讲人简介：刘旭博士是上海财经大学统计与管理学院常任教授。近年来主要研究兴趣为生成式学习、迁移学习、以及高维数据分析。获得上海市东方英才计划青年项目、上海市科学技术奖自然科学二等奖、上海市哲学社会科学优秀成果奖二等奖。在包括JASA、Biometrika、JoE、JMLR、《中国科学：数学》《统计研究》等国内外统计及相关领域的期刊发表多篇论文。主持2项国家自然科学基金面上项目、负责1项国家自然科学基金重点项目子课题，主持1项广西自然科学基金重点项目，主持1项上海市教委AI4S项目。在科学出版社出版1部专著《高维数据分析与统计推断》。
报告时间：2026年9月1日(周二)，16:30-18:00
线上地点：腾讯会议 ID：103 926 972
