Theses and Dissertations

ORCID

https://orcid.org/0009-0007-4322-2833

Advisor

Ma, Junfeng

Committee Member

Marufuzzaman, Mohammad

Committee Member

Sparrow, Michael

Committee Member

Tian, Wenmeng

Date of Degree

5-15-2026

Original embargo terms

Immediate Worldwide Access

Document Type

Dissertation - Open Access

Major

Industrial & Systems Engineering

Degree Name

Doctor of Philosophy (Ph.D.)

College

James Worth Bagley College of Engineering

Department

Department of Industrial and Systems Engineering

Abstract

This dissertation investigates the deployment of machine learning methodologies in an industrial engineering framework for the development of advanced decision support systems in the context of enrollment management. Drawing on techniques from educational data mining, the research addresses three key phases in the lifecycle of traditional and non-traditional students. First, it analyzes student retention using predictive classification models designed to identify individuals at elevated risk of attrition. Second, it employs temporal convolutional networks for time series forecasting, estimating aggregate enrollment levels over highly variable, finite planning horizons on the basis of partially observed data and using an asymmetric loss function. Third, it applies discrete-time competing risks models within a survival analysis framework to characterize longitudinal patterns of student progression, simultaneously and jointly estimating the probabilities of retention, stop-out, transfer, and graduation. Collectively, this work illustrates how contemporary predictive analytics can facilitate a shift in institutional practice from retrospective, descriptive reporting toward proactive, quantitatively rigorous decision-making.

Download

Share

COinS