Home > Events > 2017 Seminars & Colloquia > Guang Cheng, Purdue University

Guang Cheng, Purdue University

Main Content

Can we do statistical inference in a non-asymptotic way?
28 September 2017 from 4:00 PM to 5:00 PM
201 Thomas Bld
Contact Name
Lorey Burghard
Contact Phone
Add event to calendar

Many classical statistical inferential procedures are built upon large sample theory that relies on a growing amount of data information. However, in practice, it is often the case that only a small to moderate samples are available. This talk explores the possibility of establishing statistical inference with finite sample validity. The leading example is smoothing spline models under Gaussian errors. Specifically, we develop a set of non-parametric testing procedures with exact statistical guarantees in the sense that Type I and II errors are controlled for any finite sample size. An immediate  consequence of this non-asymptotic theory is a new formula (different from GCV) for selecting the optimal smoothing parameter in nonparametric testing. Simulations demonstrate that our proposed test improves over the conventional asymptotic test when sample size is small to moderate. This is an ongoing work with Yun Yang and Zuofeng Shang.


Bio: Guang Cheng is a Professor of Statistics at Purdue University.  He received his PhD in Statistics from University of Wisconsin-Madison in 2006.  His research interests include Big Data, Machine Learning and High Dimensional Statistical Inferences.  Cheng is the recipient of the NSF CAREER award, Noether Young Scholar Award and Simons Fellowship in Mathematics. Please visit his big data theory research group at 



Filed under: ,

Navigation for this Section