Sistema de Recomendação de Matrículas a partir do Perfil e do Progresso dos Estudantes no Contexto de Campus Inteligente

Abstract

In recent decades, we have achieved many technological advances and, with them, concepts of smart buildings and cities have emerged. Concomitantly, the concept of smart campus appears, understood as the use of the concepts of smart buildings and cities in a university campus. However, do not forget that the most important thing on a university campus is its community. Therefore, we have used recommender systems in a broad spectrum of applications. It is not hard to find recommenders for books, movies, music, and advertising campaigns. We focused on the community and academic processes performed daily, and this work aims to assist undergraduate students at the time of re-enrollment through recommendation systems, which direct the choice of subjects to those that best suit the student’s academic moment. To develop the recommender system, we used algorithms that analyze the student’s history, evaluate the prerequisites and equivalences attended and recommend the subjects, which are later classified and selected to generate the final list of recommendations. We used as a case study the Computer Science Course at UFPel. The recommendations were generated for the first semester of 2019 and compared with the requested, registered, and approved subjects. In addition, we developed a booster feature to reinforce recommendations for a specific period of the course curriculum. After performing the tests and comparing the students’ requested and taken, we reached more than 60% of correct answers.