KNN Algorithm for Predicting MTsN Padang Panjang Students’ to High School Placement

Education is a top priority in Indonesia’s development, with every region committed to enhancing its quality. To support individual skill development, high-quality advanced education is essential. Therefore, developing a predictive model can help students and schools prepare for the admission process to top-tier institutions. This study predicts the admission out-comes of MTsN Padang Panjang students to prestigious high schools, including MAN Insan Cendekia, SMAN 1 Sumatera Barat, SMAN 2 Padang Panjang, and SMAN 1 Padang Panjang. The model employs the K-Nearest Neighbors (KNN) algorithm utilizing historical data such as academic scores, achievement records, preferred school, and alumni distribution. The KNN algorithm was selected for its proven effectiveness in various prediction and classification tasks. Pre-processing involves managing data imbalance through SMOTE oversampling, handling outliers using the IQR method, standardizing data, and selecting relevant features. The model is evaluated using Precision, Recall, and F1-Score metrics, achieving an overall accuracy of 92%, demonstrating its effectiveness in classifying data. Unlike previous studies that focused on predicting outcomes for a single school, this study introduces a novel approach by applying a multiclass prediction model across several prestigious schools, incorporating feature selection and parameter tuning. These findings highlight the model’s potential to improve student placement predictions and aid in postsecondary school selection.