Abstract

This article presents a project of knowledge discovery in databases, which intended to help understand the influence of some factors on the academic success of undergraduate students of a Portuguese higher education institution. The steps in this process included tasks such as business understanding, data understanding, preparation of data, modelling and evaluation of results. The research contemplated predictive and descriptive activities, using the techniques of induction of decision trees and grouping. A survey, made to 86 students, included factors such as attendance in class, time of study, group study frequency, distance to the place of residence and school success. The induced induction trees revealed that the most explanatory factor of school success is the time of study and classified this success in function of the various explanatory factors. The clustering results allowed characterizing the typical student and the two or three main groups of students.

Share

COinS