Predicting Students’ Graduation at Higher Education Institutions Using Machine Learning Methods

Authors

• Jackson Gallivan, St. John Fisher University
• Alex Miller, St. John Fisher University
• Xander Ranc, St. John Fisher University
• Parker Scherlein, St. John Fisher University

Document Type

Poster Presentation

Publication Date

4-17-2026

Keywords

fsc2026

Abstract

This project aims to understand key reason(s) that university enrollees graduate or drop out by analyzing a dataset of higher level Portuguese institutions’ students and their educational outcomes. From this dataset of over 3600 different cases, and over 50 variables as predictors, we utilized machine learning techniques such as XGBoost and Random Forest with specific 80/20 training and testing splits to create predictive models of whether students would graduate or drop out. With an AUC score of 0.93 and an accuracy of 0.878, we found that the most important predictor of a student graduating is their second semester grades from their first year of University. This statistical analysis helps point administrators or advisors in the right direction when it comes to identifying at risk individuals to give extra help to in their learning experience.

Comments

Poster presented at the 2026 Fisher Showcase, St. John Fisher University, April 17, 2026.

Publication Information

Gallivan, Jackson; Miller, Alex; Ranc, Xander; and Scherlein, Parker, "Predicting Students’ Graduation at Higher Education Institutions Using Machine Learning Methods" (2026). Fisher Showcase 2026. Paper 108.
https://fisherpub.sjf.edu/fsc2026/108

Please note that the Publication Information provides general citation information and may not be appropriate for your discipline. To receive help in creating a citation based on your discipline, please visit https://libguides.sjf.edu/citations.