Naïve Bayes Classifier-Based Intelligent System for Student Academic Performance Assessment

This study proposes the design and implementation of an intelligent academic performance assessment system based on the Na ve Bayes Classifier algorithm.The system is developed using PHP and MySQL and is architected with a lightweight and modular structure tailored for school environments with limited technological resources.It integrates a classification engine capable of processing diverse academic indicators, including test scores, attendance records, and behavioral data, to classify students into performance categories (low, medium, high).Compared to conventional assessment methods that often rely on manual judgment and are prone to inconsistency, this system offers a data-driven and objective alternative that supports evidence-based educational decision-making.Across ten testing iterations, the system achieved an average classification accuracy of 96.67%, demonstrating its predictive reliability.Moreover, user evaluations involving 230 respondents (teachers and students) reported an overall satisfaction rate of 86.8%, indicating strong acceptance in terms of usability and effectiveness.The study highlights the advantages of Na ve Bayes over more complex algorithms such as XGBoost and neural networks, emphasizing its ease of interpretation, computational efficiency, and practical deployability in real-world educational contexts.The system's predictive outputs enable early identification of students requiring academic intervention and support differentiated instruction, ultimately contributing to the enhancement of personalized learning pathways.These findings reinforce the role of machine learning, particularly interpretable and resource-efficient models, in transforming traditional assessments into intelligent and scalable educational solutions.