Title: Intelligent data mining for real-time academic insights to drive proactive decision-making in higher education innovation

Authors: K. Shilpa; D. Suresha

Addresses: Department of Computer Science, Institute of Computer Science and Information Science, Srinivas University, Mukka, Mangalore, India ' Department of Computer Science and Engineering, Srinivas Institute of Technology, Institute of Computer Science and Information Science, Srinivas University, Mukka, Mangalore, India

Abstract: Advanced real-time academic insights require complex data mining tools because higher education student performance forecasting is challenging. The project utilises machine learning and educational data mining to develop a data-driven framework that enhances higher education student outcomes and reduces dropout rates. The study uses over 4000 undergraduate and postgraduate student records from Bangalore universities to combine academic accomplishment, staff involvement, attendance, engagement and socioeconomic variables. Seven supervised learning algorithms were tested for their accuracy in predicting student academic success and dropout likelihood using the KDD paradigm. The framework includes DT, RF, SVM, kNN, LR, NB and MLR. Random Forest had the highest prediction accuracy for real-time student monitoring and intervention planning (91.63%) with a robust F1-score (0.9332) and AUC-ROC (0.8748). In-depth correlation, feature importance analysis, clustering, sentiment mining and sentiment analysis revealed that study hours, faculty support and attendance significantly impact academic achievement. Stress was negatively correlated. Risk reduction and instructional approaches were also provided via hierarchical clustering and tiered performance analysis. These results demonstrate how current data mining can enhance personalised learning, retention and informed decision-making in proactive education.

Keywords: academic performance prediction; decision trees; random forests; SVM; support vector machines; k-nearest neighbours; logistic regression; Naïve Bayes; MLR; multiple linear regression; KDD; knowledge discovery in databases.

DOI: 10.1504/IJMIE.2026.154569

International Journal of Management in Education, 2026 Vol.20 No.4, pp.407 - 440

Received: 13 Mar 2025
Accepted: 15 Jun 2025
Published online: 06 Jul 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article