Mineração de Padrões Sequenciais de Sentimentos: Um Estudo de Caso na Detecção de Propensão à Evasão Escolar na Educação Superior

Resumo

In view of the high dropout rates of Brazilian undergraduate courses, this paper investigates the hypothesis that the use of the underlying sentiment in the student-university interactions can improve sequential pattern mining based dropout prediction. To this end, the present work applied an adapted version of the method proposed in (Removed for blind review) to a set of historical data of a university. Quantitative results of the experiments confirmed the hypothesis raised. In addition, the patterns discovered led to a set of actions that can be added to the universitys dropout combat program.