Prediction of Science and Social Science Students Interests Using Decision Tree Algorithm with CRISP-DM

Selecting a major in Senior High School significantly shapes students’ academic paths and future careers. However, the current process often lacks objectivity, relying on subjective teacher consultations and academic data without standardized analysis. This study addresses this gap by developing a decision support system using the Decision Tree algorithm to assist students at Al-Istiqomah High School in choosing between science and social science majors, based on academic performance and non-academic factors like attitudes and attendance. The study follows the CRISP-DM methodology, which includes six stages: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Academic records from 123 students were used to build the model. The Decision Tree algorithm identified mathematics, biology, and physics scores as key predictors for major classification. The model demonstrated high predictive performance, with 96% accuracy, 100% precision, and 95% recall. Additionally, an Area Under the Curve (AUC) of 97% confirmed the model’s robust ability to distinguish between science and social science tracks. This system was implemented as a user-friendly web application using Streamlit, enabling students and educators to input data and receive immediate major predictions. By offering objective, data-driven recommendations, the system helps students make more informed decisions about their academic futures and provides educators with targeted, evidence-based advice. These results highlight the Decision Tree algorithm as an effective, efficient, and practical tool for enhancing the academic advising process and supporting students in selecting the major that best fits their strengths and interests.