USE OF MACHINE LEARNING ALGORITHMS FOR UNIVERSITY MANAGEMENT: A SYSTEMATIC REVIEW

Abstract

This research aims to carry out a systematic review on the application of Machine Learning (ML) in university management. The study allows us to analyze previous research on the use of the most used ML algorithms and highlighted as the best, for a discussion and agenda. The searches were made in the international scientific databases Scopus and Web of Science, using the keywords Machine Learning, University, Higher Education. A total of 32 articles were selected for the sample. The results indicate that most surveys use more than one ML algorithm to perform predictions and that the Support Vector Machine (SVM) is the algorithm highlighted in most surveys as the best performer. Another conclusion identified is that most of the articles assess the risk of school dropout, student academic performance, predict student results, and analyze dropout in order to avoid dropout, dropout, or attrition from courses by the student.