ABSTRACT
This paper tackles the problem of dropout of undergraduate students
in a private university, by using Educational Data Mining
(EDM) techniques. The EDM is an emerging area, concerned with
developing methods for exploring the increasingly large-scale data
that come from educational settings and using those methods to
better understand students and the settings which they learn in. In
this work, EDM is used to identify profiles of students who withdraw
from their engineering courses. The considered dataset is
composed of 53 attributes, involving financial and academic aspects
of 2,925 engineering students. Preliminary results have identified
some attributes that are related to the dropout in engineering courses,
such as: the semester of the year (students are more prone to
dropout in the first half of the year), attendance, grades (in this
case median is more important than the mean value) and number
of credits in the previous semester, and the current semester the
student is enrolled (students bellow the 5th semester have a higher
tendency to dropout).