Student performance prediction employing k-Nearest Neighbor Classification model and meta-heuristic algorithms

The precision of predicting student performance in Mathematics is pivotal for educational advancement, relying heavily on advanced machine learning (ML) methodologies. This predictive approach involves analyzing comprehensive datasets, emphasizing academic records, demographics, and various educational metrics, with a particular focus on Mathematics. Techniques like classification, regression analysis, decision trees, and neural networks yield highly accurate projections, aiding in timely interventions crucial for supporting students navigating mathematical complexities. These algorithms optimize resource allocation within educational institutions, identifying and aiding students requiring extra assistance in their mathematical pursuits. This study pioneers the enhancement of predictive capabilities in Mathematics through the innovative integration of the K-Nearest Neighbor Classification (KNNC) model with 2 novel optimization techniques: the Honey Badger Algorithm (HBA) and the Arithmetic Optimization Algorithm (AOA). By harnessing these cutting-edge ML and bio-inspired algorithms, the research is dedicated to pushing the boundaries of precision and reliability in forecasting, with a specific focus on elevating educational outcomes within the domain of Mathematics. The outcomes obtained for G1 and G3 reveal that the KNHB model exhibited outstanding performance in predicting and categorizing G3. It achieved remarkable Accuracy and Precision of 0.921 and 0.92, respectively. Moreover, the KNHB proved to be the most precise predictor for G1 value prediction, with Accuracy and Precision scores of 0.899% and 0.9%, respectively, in the prediction task.