报告人 ：蒋建成（University of North Carolina at Charlotte）

时间：2023年06月08日 14:30--

地址：理科楼LA104

摘要：We propose an empirical likelihood ratio (ELR) test for comparing any two supervised learning models, which may be nested, non-nested, overlapping, mis-specified, or correctly specified. The test compares the prediction losses of models based on the cross-validation. and allows for heteroscedasticity of the errors. We establish asymptotic null and alternative distributions of the ELR test for comparing two nonparametric learning models under a general framework of convex loss functions. However, the prediction losses from the cross-validation involve repeatedly fitting the models with one observation left out, which  leads to a heavy computational burden. We introduce an easy-to-implement ELR test which requires fitting the models only once and shares the same asymptotics as the original one. The proposed tests are applied to compare additive models with varying-coefficient models. Furthermore, a scalable distributed ELR test is proposed for testing the importance of a group of variables in possibly mis-specified additive models with massive data. It is shown that the distributed ELR performs the same as the ideal ELR with full data running on one machine. Simulations show that the proposed tests work well and have favorable finite-sample performance over some existing approaches. The methodology is validated in an empirical application.

简介：Dr. Jiancheng Jiang is a professor of statistics at the University of North Carolina at Charlotte, USA. He was appointed as chair professor of Nankai University in 2017-2020 and served as the statistics program coordinator at UNC Charlotte and the associate editor of Statistica Sinica and other journals since 2017. He has been awarded several NSF/NIH grants since 2004. His research interest ranges from (bio)statistics to econometrics and data science.

邀请人：张志民

欢迎广大师生积极参与！