Abstract: Artificial Intelligence (AI) has become a potent instrument in education, facilitating predictive modeling, adaptive learning, and early recognition of student at-risk status. Conventional means of performance assessment—e.g., tests and teacher ratings—are typically reactive, giving feedback only after students have performed below potential. This paper discusses the use of AI methodologies to predict student academic performance on the basis of several parameters such as attendance, assignment grades, and previous grades. Three machine learning models—Decision Tree, Random Forest, and Neural Network—were tested and compared on a student dataset. The models were compared on a variety of performance metrics such as accuracy, precision, recall, and F1-score to cover all aspects of evaluation. Experimental findings indicate that the Neural Network performed better than the Decision Tree and Random Forest, with the highest rate of accuracy in forecasting student outcomes. Such findings propose that AI is able to improve proactive education initiatives immensely by foreseeing struggling students early and facilitating customized learning interventions. The research concludes that the incorporation of AI-based predictive systems in academic institutions is capable of revolutionizing the evaluation process from reactive measurement to proactive student intervention.

Conventional evaluation tools—like tests, manual grading, and teacher observation—are usually reactive in nature, recognizing performance problems only after students underperform. AI can, on the other hand, change the paradigm of education to proactive and preventive by projecting outcomes beforehand and suggesting individualized interventions.
This work investigates the use of AI methods for the prediction of student academic performance based on attendance, assignment grades, previous grades, and learning activity as input features of primary importance. Three machine learning algorithms—Decision Tree, Random Forest, and Neural Network—were trained and tested on student datasets. Performance was compared on several measures, such as accuracy, precision, recall, and F1-score, to provide a thorough evaluation.

Experimental findings reveal that the Neural Network outperformed Decision Tree and Random Forest models consistently, yielding the maximum accuracy for predicting outcomes. The results show the promise of AI-based prediction systems to revolutionize pedagogy. With the ability to initiate interventions early on, AI can address issues related to low dropout rates, enhance student motivation, and increase academic achievement. The research infers that the use of AI-based predictive systems by institutions can aid in more adaptive, data-driven, and student-centric education.

Downloads: | DOI: 10.17148/IJIREEICE.2025.13918