Academic performance profiles : an intelligent predictive model based on data mining

Abstract

It is well known that academic achievement is one of the key aspects in the development of educational activities and it strongly determines the chances of success during and after a university career. It is therefore important to try and effectively monitor students’ performance in order to prevent problems from emerging, as well as, to be able to provide academic coaching when the performance is not adequate. The aforementioned problem-anticipation possibility is closely related to the ability to predict the most probable situation based on concrete information. In an academic achievement framework, it is desirable to be able to predict students’ performance considering concrete individual parameters. This work outlines the results obtained by an academicachievement prediction model based on data mining algorithms which uses socioeconomic information as well as, students’ grades. The tests were carried out at National Technological University, Resistencia Regional Faculty (UTN-FRRe), during the AED-Algoritmos y Estructuras de Datos (Algorithms and Data Structures) class throughout the 2013, 2014, 2015 and 2016 terms. The results obtained confirmed adequate behaviour of the model which has been validated for both description and prediction of academic achievement profiles.