Work

Student Education Performance and Behavioral Patterns using Machine Learning Techniques

Year: 2025

Type: conference-paper

Abstract: Student dropout remains a critical concern in higher education, affecting not only individual learners but also institutions and society at large. With the increasing availability of educational data online, predictive analytics has emerged as a valuable tool to address this issue by identifying students who may be at risk of leaving their studies prematurely. This study aims to develop and evaluate predictive models that can accurately predict student dropout based on academic performance, demographic information, and behavioral indicators. Using a dataset from a higher education institution collected in 2021, we carried out data preprocessing and model training using various machine learning techniques such as Random Forest, Logistic Regression, XGBoost, SVM, and more. However, many existing models face challenges related to limited generalizability and the lack of emotional and sentimental features. By addressing some of these limitations, this report explores the strengths and weaknesses of each model built and compares their predictive accuracy to identify the most effective one. Ultimately, this research is aimed to support early intervention strategies to enable educators to provide timely assistance to at-risk students and contribute to improved retention rates in higher education. The findings also offer insights for researchers and policymakers seeking to enhance predictive systems and student support mechanisms.

Source: IEEE Connecting the Unconnected Europe, Middle East, and Africa Regional Summit (CTUS-EMEA)

DOI: 10.1109/ctus-emea68989.2025.11592196