A Review of IA Use in Education Analysis

• Main Authors: Magalhaes dos Santos, Joao Kleber, da Rocha, Henrique Oliveira, Rodrigues Okamura, Erick Massaru, Araujo Dias, Victor Augusto, Pessoa de Melo, Lucas Hipolito, Oliveira Viana, Giovani Braga, Rodrigues, Vinicius Cardoso, da Silva, Daniel Alves
• Format: Conference Proceeding
• Language: English
• Subjects: Analytical models; Dropout; Education; Machine learning; Neural networks; Pandemics; Prediction; Psychology; Statistical analysis; Students; University
• ISSN: 2768-0045
• DOI: 10.1109/WCNPS60622.2023.10344527

Higher education evasion is a problem that is present worldwide, and that has many consequences to the society as a whole. So, given the importance of this problem, a lot of literature, in the areas of computer science, psychology and statistics, was produced with the intention of predicting and trying to mitigate dropout rates in higher education. Aiming to address this problem this work makes an analysis of the literature about the topic and produce a statistical analysis of a dataset provided by INEP (National Institute of Educational Studies and Research Anísio Teixeira) that brings indicators of dropout of undergraduate courses from universities in Brazil, and a comparison was made with specific data from engineering courses at UNB using two machine learning methods: neural networks and regression. Observing the results of the model made, they show that the model could not surpass the models in the literature, however the analysis carried out revealed interesting data on dropout rates in Brazilian higher education and it is also noticed that dropout increased considerably in the pandemic period. In view of this, predictive models are very useful methods for identifying students at risk of dropout