Predição do rendimento dos alunos em lógica de programação com base no desempenho das disciplinas do primeiro período do curso de ciências e tecnologia utilizando técnicas de mineração de dados

Abstract

The high rates of university student disapproval and dropout in the initial courses of programming present a worrying statistics faced by the coordinators of the Technology programs. The problem of students disapproval is often pointed as an influential factor in dropping out of university programs. This work proposes the use of techniques of Educational Data Mining to predict the performance of students in the course of Programming Logic, of the second period of the Bachelor of Science and Technology program at UFRN, based on performance in the courses of the first period of that program. The results showed that it is possible to infer students performance with an accuracy of up to 77%, this information being useful for planning actions to avoid disapproval/dropout and, especially, to personalize the teaching of programming logic.