Técnica de clusterização aplicada à análise de perfis socioeconômicos de estudantes concluintes de cursos de computação

Abstract

The different social and economic classes of undergraduate students can impact the course of academic training and the permanence of such students in Brazilian higher education institutions. This course conclusion work applied a data mining technique called K-means clustering to the microdata of the 2017 National Student Performance Exam (ENADE), an exam applied by the National Institute of Educational Studies and Research Anísio Teixeira (Inep), with the aim of analyzing the contexts that separate graduates from the various computer courses, whether bachelor’s or licentiate, using socioeconomic data. The results pointed to four large groups of students and, based on their analysis, it is possible to list a profile of a graduate student of computing in the year analyzed, since the clusters have several characteristics in common, such as: most students are of the sex male, single, white, opted for the face-to-face modality, attended high school in public schools, among others. However, some characteristics were found in specific groups, for example there is a group of graduates who are from full-time public institutions.