Ph.D. Student Yixuan Qiu wins Third Place at ICSA Midwest Chapter Meeting's Poster Award - Department of Statistics - Purdue University Skip to main content

Ph.D. Student Yixuan Qiu wins Third Place at ICSA Midwest Chapter Meeting's Poster Award

11-13-2015

Ph.D. Student Yixuan Qiu wins third place award at the ICSA Midwest Chapter meeting’s student poster competition on October 26, 2015.

Congratulations to Ph.D. student Yixuan Qiu (advisors Dr. Lingsong Zhang and Dr. Xiao Wang), who was recently awarded the third place prize for the student poster competition at the International Chinese Statistical Association (ICSA) Midwest Chapter meeting for his poster titled Semi-Bayes Problems and Methods co-authored with Professors Chuanhai Liu and Lingsong Zhang.

This award is bestowed and evaluated to the students with the best presence and quality of statistical significance, novelty of the technical content, accessibility to a broad audience, accuracy of language, layout of content, readability of the presentation, and graphics. The top three posters were awarded.

Yixuan’s poster focused on developing exact inference for problems when prior information is partially given, which are called Semi-Bayes problems. The Semi-Bayes approach is a new framework for statistical inference. It generalizes the traditional Bayes paradigm in a way that the prior distribution can be only partially specified. For example, we can allow unknown hyper-parameters in the prior. Based on the Inferential Models theory developed by Prof Liu and his coauthors, the Semi-Bayes approach is able to draw exact and efficient inference with partial information on the prior distribution, and the result does not rely on large sample approximations. We compare Semi-Bayes with other existing approaches such as Empirical Bayes in a number of interval estimation problems, and demonstrate that the Semi-Bayes approach is able to preserve the nominal coverage rate in any finite sample scenario. This result indicates that Semi-Bayes is especially suitable for real problems in which sample size is only finite and Type I error needs to be strictly controlled.  

Join us in congratulating Yixuan on his award.    

Purdue Department of Statistics, 150 N. University St, West Lafayette, IN 47907

Phone: (765) 494-6030, Fax: (765) 494-0558

© 2024 Purdue University | An equal access/equal opportunity university | Copyright Complaints | DOE Degree Scorecards

Trouble with this page? Accessibility issues? Please contact the College of Science.