Application of Machine Learning for Predicting SNBP Graduation of High School Students in Kabupaten Toba

Education plays a crucial role in enhancing human resources and national competitiveness. In Kabupaten Toba, North Sumatra, top-performing high schools such as SMAN 1 Balige and SMAN 2 Balige have maintained strong academic reputations with accreditation A. However, their students' success rate in the National Selection Based on Achievement (SNBP) for public university admission remains low. In 2023, only 42 out of 309 students passed SNBP, while 267 students failed. This study aims to address one of the key factors behind this low success rate-students' difficulty in selecting a university major aligned with their academic strengths and interests. Many students rely on subjective preferences, overlooking important academic and competitive considerations. To support more informed decision-making, this research applies machine learning techniques to predict students' chances of SNBP success based on historical academic data. Decision Tree and Random Forest algorithms were chosen for their interpretability and effectiveness in managing complex datasets. Hyperparameter tuning was also applied to optimize model performance and highlight the most influential variables. Following the CRISP-DM methodology encompassing business understanding, data understanding, data preparation, modeling, and evaluation the study uses academic records over three years along with SNBP outcomes. The resulting predictive model is expected to guide students in selecting appropriate majors, thereby improving their SNBP success rates. Additionally, the model can assist schools in refining academic counseling strategies and enhancing overall education quality.