Improvement of k-NN Algorithm Performance in Classifying High School Student's Majors

Classifying high school students into appropriate academic majors is essential to align their academic strengths with future studies and career paths. This study aims to improve the accuracy of classifying students' majors into Mathematics, Language, and Science using the k-Nearest Neighbors (k-NN) algorithm. The dataset used initially had 243 records because it was still imbalanced. Then, using the SMOTE method, it became balanced with 304 records. The training and testing data separation was 70:30, and the optimized K value was 9. The model achieved an accuracy of 94.51%. The results show that with careful tuning, k-NN is a highly effective method for classifying students into appropriate academic tracks. This study demonstrates the potential of k-NN as a decision support tool for educational institutions, offering a data-driven approach to assist the student major placement process. Future studies can explore additional features and alternative machine learning techniques to further improve the model's performance.