Uma abordagem de aprendizado de máquina para o auxílio no diagnóstico de evasão escolar em turmas do ensino médio

Abstract

School dropout is one of the major problems faced by education systems in many countries around the world. There are several studies that evaluate possible causes for this situation, such as socio-economic profile of the student, characteristics and structure of the school and also the region where it is inserted, profile of teachers, as well as several others. However, much of this information is imprecise or not known for a number of reasons that go beyond the capacity of educational management systems. Therefore, the need to use systems that automate the analysis of possible causes is extremely important for the decision-making process to be taken by managers. The purpose of this paper is to apply machine learning techniques to perform this type of analysis. The advent of machine learning techniques makes the optimal exploitation of databases possible, allowing the emergence of new perspectives and information that may aid in the decision-making process. Thus, it is possible to predict possible high school dropout with relative accuracy, based on determinant factors found in data from previous academic years as family socioeconomic information. For the accomplishment of the experiments we used a database obtained from the reality of high school classes of schools in Portugal.