Chrun risk mitigation models for student’s behavior

In order to prevent the phenomenon of abandonment (churn), the objective of this paper is to analyze university student’s careers. The results of our research will be used to plan activities of collective or single tutoring and to measure the efficiency and efficacy for specific courses. In particular, the analysis considers years of history for two faculties in the University of Pavia: psychology and biology. In order to estimate for each student a measure of churn risk, our methodological approach is based on a duration predictive model. Empirical evidences are given on the basis of a real data set.