IEOR Seminar Series: Pengyi Shi, Purdue University

Event details Date: Monday, August 31, 2026 Time: 3:30 PM to 3:30 PM Location: 3108 Etcheverry Hall Type: Lecture / Workshop Audience: Alumni,Faculty,Prospective Students,Public,Staff,Students About this event Title: Admission Decisions under Imperfect Classification Abstract: Many capacity-constrained service systems use machine-learning predictions to prioritize individuals for admission. Yet better predictive accuracy does not necessarily lead to better operational decisions. We study a general class of priority-based admission-control problems under imperfect classification, motivated by community-based incarceration diversion programs. We first develop a decision-aligned ranking metric that connects predictive models to system-level allocation cost. We then develop an uncertainty-quantification framework for individual-level decision, which translates estimation uncertainty into ranking and admission-boundary uncertainty. Using data across seven Illinois counties in the Adult Redeploy Illinois (ARI) program, we calibrate a semi-synthetic case study to examine how alternative ML algorithms and model-selection criteria affect priority rankings, allocation costs, and uncertainty in individual admission decisions. This is a joint work with Zhiqiang Zhang and Amy Ward at the University of Chicago Booth School of Business. Bio: Pengyi Shi is an associate professor at the Daniels School of Business, Purdue University. She received her Ph.D. degree in Industrial Engineering from Georgia Institute of Technology before joining Purdue in 2014. Her research interests include data-driven modeling and decision-making in healthcare operations. She has collaborated with practitioners from different healthcare organizations worldwide. Most recently, she is collaborating with community correction programs to develop data-based evaluation and human-in-the-loop machine learning algorithms. Her research has won the first place of IISE Outstanding Innovation in Service Engineering Award in 2023, the first place of MSOM Responsible Research in OM Award in 2021, the first place of Pierskalla Best Paper Award in 2018, and a finalist of the... Official event details: https://events.berkeley.edu/IEOR/event/327954-ieor-seminar-series-pengyi-shi-purdue-university