An Interpretable Machine Learning Application for Predicting and Improving University Readiness

Supporting high school students in making informed decisions about their academic paths is crucial for reducing dropout rates and improving university success. This paper presents a web-based application built using data from 126,577 students across six large Hungarian universities. Using predictive analytics and interpretable machine learning techniques, such as SHapley Additive exPlanations (SHAP), the system predicts students' university performance and identifies skills and subjects that would most effectively boost their academic success. By providing actionable insights and personalized guidance, this dual-purpose application supports high school students in choosing suitable majors and preparing effectively for university challenges, ultimately contributing to better academic outcomes.