Sistema de recomendação de instituição de ensino superior para atenuar evasão utilizando análise de perfis de egressos

Abstract

Evasion in higher education is a social, academic and economic waste. Studies have been carried out in order to identify the coefficients that lead to evasion and, likewise, verify how to oppose such coefficients. This work proposes the use of methods for Knowledge-Discovery in Databases and Recommendation System in order to suggest educational institutions to future students. The first process performed consisted of understanding the database of gls ENADE, cleaning, pre-processing, transforming the data and performing Data Mining tasks to obtain profiles of the graduates. The Data Mining tasks used covered techniques of clustering and classification methods to define the profiles. The second process performed was the development of a Recommendation System, using a textit Web application to collect user data in order to generate a profile. In this system, the user’s profile is compared with the profiles extracted in the first process in order to identify the similarity between them. Subsequently, the system recommends educational institutions that have a profile more similar to the user’s profile.