报告人:夏应存 （电子科技大学）

日 期:2019年6月18日

时 间:19:30

地 点:理科楼 LA106

摘 要: Non-negative Matrix Factorization (NMF) has been used in many disciplines of research, where the number of factors plays a crucial role. However, a fully data-driven method for selecting the number of factors is yet not available. Based on the fact that the appropriate number of factors should generate the best prediction, in this paper we propose a selection method using a two-step delete-one-out approach, called twice cross-validation (TCV). Intensive simulations suggest that the proposed method works well especially when the number of factors is much less than the dimension of variables and meanwhile the sample size is reasonably big. As an important application of NMF, the source apportionment of air pollution is considered for data from Singapore and other places.

报告人简介：夏应存，电子科技大学教授，1999年获得香港大学统计学博士学位，2000年至2001年在伦敦政治经济学院统计系任助理研究员，2000年至2003 年为剑桥大学动物学系助理研究员，2003至今在新加坡国立大学统计与应用概率系工作。现任Annals of Statistics, Computational Statistics and Data Analysis等期刊的副主编。夏应存的研究兴趣包括非参数回归分析，计量经济及金融时间序列分析，高维数据分析，环境与健康的统计分析。在学术期刊PNAS， AoS， JASA， JESSB， American Naturist， JoE 等发表多篇论文。部分论文在JRSSB，Statistical Sciences及statistica Sinica等期刊上专题讨论；Nature News等十几家学术媒体对其工作做过专题报道。

公司联系人:杨 虎

欢迎广大师生积极参与！