Prediction of academic performance using artificial intelligence techniques

Abstract

The aim of this article is to predict the academic performance of higher education students, considering several influential factors, applying artificial intelligence techniques (classifiers). Although such factors have been widely analyzed from quantitative and qualitative approaches, they still represent research opportunities using artificial intelligence tools, particularly in academic performance prediction. With the definition of influential factors (educational, family background, social and economics, habits and customs, among others), a methodology was designed to train a system able to a priori classify a new student, in one of the five categories of academic performance. This classification allows an educational institution to have an early identification of students with potential academic performance problems. From this knowledge the institution can deploy immediate mitigation action. The methodology was applied to a sample of students from a public university in Colombia, obtaining a success level of 91.7%.