Aplicação de técnicas de mineração de dados e Learning Analytics para predição de evasão de alunos nos cursos de Ciência da Computação e Engenharias da UFPel

Abstract

Educational Management Systems store a large amount of data from interaction of not only students and professors but also of students and the educational environment. Analyze and find patterns manually from a huge amount of data is hard, so Educational Data Mining (EDM) is widely used. This work presents a model that can predict the student’s risk of dropout using data from the first three semesters attended by Computer Science (N=1516) and Engineering (N=6351) Undergraduate students from UFPel. Both data sets used the same attributes, which was a total of 22 attributes between socio-economic and academic. This work uses the CRISP-DM methodology e data from UFPel Management System (called Cobalto). The results are shown for five algorithms. For Computer Science, the Logistic Regression algorithm a precision of 91.24% and a Recall of 92.17% is presented. For Engineering, the Random Forest algorithm a precision of 83.40% and a Recall of 79.48% is presented. For both data bases (Computer Science and Engineering), the results indicate that it is possible to use a prediction model using only the data from the first three semesters of the course.