Data Mining Model Approach for Employment Prediction for University Graduates

Abstract

Graduate employability prediction has become increasingly important as universities seek to understand the factors influencing employment outcomes and to improve academic planning. However, prior studies in this area often rely on limited datasets, evaluate only a narrow range of models, and lack systematic feature assessment, which restricts the robustness and generalizability of their findings. This study addresses these limitations by developing a comprehensive multi-model data-mining framework to predict the employability of graduates from Rajamangala University of Technology Lanna (RMUTL). A dataset of 4,352 graduate records from the 2023 academic year was analyzed. Three filter-based featureselection techniques—chi-square, information gain, and correlation-based evaluation—were applied to identify the most influential predictors. Five machine-learning algorithms (Decision Tree, Random Forest, Gradient Boosted Trees, Naïve Bayes, and K-Nearest Neighbors) were trained and evaluated using accuracy, precision, recall, F1-score, and AUC. The results show that Random Forest achieved the highest accuracy (83.72%), while Gradient Boosted Trees yielded the highest AUC (0.813), indicating superior class-separation performance. Key predictive factors identified across models included curriculum, education level, department, faculty, campus, gender, and GPA level. This study provides a structured comparative modeling framework and identifies institution-specific predictors that influence graduate employability. The findings offer practical implications for curriculum enhancement, evidence-based academic planning, and career-guidance development aimed at improving employment outcomes for RMUTL graduates.