Machine Learning-Based Identification of Students at Risk of Poor Academic Performance

ATHUL VARGHESE

Machine Learning-Based Identification of Students at Risk of Poor Academic Performance

Author(s)	Mr. ATHUL VARGHESE
Country	India
Abstract	Identifying students who are at risk of poor academic performance can help educational institutions provide timely academic support and intervention. This study investigates whether machine learning can identify students at risk using demographic, social, behavioral, and educational characteristics. The study uses the UCI Student Performance mathematics dataset containing 395 students. The final mathematics grade (G3) was converted into a binary target: students with G3 below 10 were classified as at risk, while students with G3 of 10 or above were classified as not at risk. First- and second-period grades (G1 and G2) were excluded to avoid direct reliance on previous examination performance and to examine a more useful early-warning setting. Five classification algorithms—Logistic Regression, Support Vector Machine, Random Forest, Decision Tree, and K-Nearest Neighbors—were evaluated using stratified five-fold cross validation. Accuracy, precision, recall, F1-score, and ROC-AUC were used as evaluation measures, while confusion-matrix analysis was used to examine classification errors. Random Forest achieved the highest accuracy (69.11%) and ROC-AUC (69.00%), whereas Logistic Regression produced the highest recall (56.15%) and F1-score (51.23%) among the tested models. Feature analysis identified school support, sex, previous failures, social activity, and absenteeism among the more influential variables. The findings indicate that machine learning can identify meaningful patterns associated with academic risk, although predictive performance remains moderate and false predictions are substantial. The results also demonstrate that accuracy alone can be misleading when the practical objective is to identify at-risk students. Machine learning may therefore be useful as a supplementary early-warning tool, provided that predictions are interpreted alongside educator judgment, student context, and appropriate academic-support procedures rather than used for automatic decisions.
Keywords	Machine Learning,Student Performance,Academic Risk,Educational Data Mining,Predictive Analytics
Field	Computer > Data / Information
Published In	Volume 7, Issue 5, September-October 2026
Published On	2026-09-07

View / Download PDF File

E-ISSN 3048-7641

doi

CrossRef DOI is assigned to each research paper published in our journal.

AIJFR DOI prefix is
10.63363/aijfr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 7 Isu 4 Cover Page Vol 7 Isu 3 Cover Page Vol 7 Isu 2

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About AIJFR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9898-79-39-59	editor@aijfr.com

Advanced International Journal for Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Machine Learning-Based Identification of Students at Risk of Poor Academic Performance

Share this