勷勤数学•领军学者报告

题      目：Finite Mixture of Regression Models and Biomedical Applications

报  告  人：林诗丽   教授  (邀请人：吴琴 )

                                     Ohio State University

时      间：6月20日  16：00-17：00

地     点：数科院东楼阶梯二楼报告厅

报告人简介:

       Shili Lin is a Professor of Statistics at the Ohio State University, the Chair of the University-wise Interdisciplinary Ph.D. Program in Biostatistics, and the Founding Director of the Computational Health and Life Sciences Community of Practice at the Translational Data Analytics Institute. She serves in the Board of Directors of the Canadian Statistical Sciences Institute, among many other national and international committees. She is currently serving or has served as an Associate Editor in several journals, including the Journal of the American Statistical Association and Biometrics. She is a Fellow of the American Statistical Association, a Fellow of the Institute of Mathematical Statistics, a Fellow of the American Association for the Advancement of Science, and an elected member of the International Statistical Institute. The international conference “Advances in Statistical and Computational Methods for Analysis of Biomedical, Genetic, and Omics Data” was held in March 2023 to honor Shili Lin for her research, mentoring, and leadership.

摘      要：

       Finite mixture models are ubiquitous in mathematical and statistical theory, methodology, and applications in many areas of scientific inquiries. Among them is a class of models that are generally referred to as finite mixture of regression models, where each component of the mixture is influenced by a set of covariates, a setting frequently encountered and particularly relevant and important in the big data era. The high dimensionality complicates the problem significantly, leading to numerous new developments of theory and methods for sparsity treatments. In this talk, I will touch on a number of issues in finite mixture of regression models, including (1) feature selection in a sparse model with high-dimensional feature space, (2) regularization with diverging number of parameters, (3) heterogeneity of covariate effects in the presence of censoring, and (4) semi-parametric finite mixture of varying coefficient models. Multiple applications in biomedical and genomics research will be used to motivate the problems and illustrate the utilities of the solutions discussed.

          欢迎老师、同学们参加、交流！