Modeling the International Mobility of Tec Graduates: Predicting Emigration with Machine Learning

This study aims to develop and validate a data-driven predictive model that estimate the probability of Tecnológico de Monterrey graduates emigrating abroad within five to ten years after graduation, thereby generating actionable insights to strengthen the university¿s internationalization strategy and alumni follow-up programs. The dataset consists of demographic, academic, and institutional variables, including socio-family background, professional trajectory, and well-being. A calibrated CatBoost model was developed to estimate the probability of student emigration, highlighting its balanced performance. The CatBoost model achieved balanced performance (77 % accuracy), while highlighting nine high-leverage characteristics (campus, program, age, parental education, scholarships, and postgraduate status). This calibrated pipeline can shift graduate mobility management from reactive tracking to proactive, evidence-based engagement, preserving alumni ties while maximizing Tec¿s global impact. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Modeling the International Mobility of Tec Graduates: Predicting Emigration with Machine Learning Chapter in Scopus