Modelos de riscos competitivos no estudo de evasão discente

Abstract

Dropout is a problem faced by public and private universities. Therefore, the motivation of this work is to investigate the characteristics of students enrolled in a regular undergraduate course and verify which one is related with dropout and/or graduate. For the study we consider competitive risk models, and we propose to use the transformation of complementary log-log for the risk function into discrete data to study specific cause risk at a period of time (calculated in semesters). For the parameters estimations we used Bayesian Inference. We verified that the model proposed in this work can be adjusted to the theoretical data. In real data from the Applied Mathematics course to the Institute of Mathematical and Computer Sciences (ICMC), we concluded that parental education and admission type influence in dropout outcome