Machine Learning for the Identification of Students at Risk of Academic Desertion

• Published in: Learning Technology for Education Challenges Vol. 1011; pp. 462 - 473
• Main Authors: Zea, Leidy Daniela Forero, Reina, Yudy Fernanda Piñeros, Molano, José Ignacio Rodríguez
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2019; Springer International Publishing
• Series: Communications in Computer and Information Science
• Subjects: Academic desertion; Algorithm; Artificial intelligence; Machine Learning
• ISBN: 9783030207977; 3030207978
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-030-20798-4_40

In Latin America, desertion rates in higher education range between 40% and 75%. There are many reasons for a student to deserted of their studies. However, the importance of identifying the level of risk related to such desertion is reflected in the socio-economic impact for the institutions as well as for the country. Technological advancements in database management and artificial intelligence have led to the development of techniques such as Machine Learning, which supports decision-making when facing a problem and adapts accordingly to the required conditions. The following article shows a case study of the identification of students in Industrial Engineering at risk of dropping out in the Universidad Distrital Francisco José de Caldas from the 2003-1 to 2018-1 academic semesters. The algorithm is selected based on which is more suitable to the nature of data, through the comparison of automated learning techniques in Azure Machine Learning Studio.