Uso de análise de sobrevivência como ferramenta na obtenção de indicadores de permanência no ensino superior

Abstract

This paper proposes an integrated approach of statistical techniques, business intelligence and data science concepts to create a survival analysis model aimed at understanding student retention in higher education. Using computer science courses as a case study, different groups are compared in order to calculate the probability of students staying until the end of the course. The work makes use of the SABIA platform to support data-driven management, highlighting the importance of technological tools in academic analysis. The results show patterns between courses in the same area when considering dropout as an event of interest, with steeper drops in the initial periods, especially at critical moments, such as the second semester, which show lower probabilities of permanence compared to the first semester, as well as the Other entry modality, which in the final periods reflects the occurrence of dropout, highlighting the importance of personalised interventions to prevent dropout and waste of resources, contributing to more efficient and effective management of higher education institutions.