Evaluation of algorithms to predict graduation rate in higher education institutions by applying educational data mining
Por favor, use este identificador para citar o enlazar este ítem:
http://hdl.handle.net/10045/93995
Título: | Evaluation of algorithms to predict graduation rate in higher education institutions by applying educational data mining |
---|---|
Autor/es: | Moscoso-Zea, Oswaldo | Saa, Pablo | Luján-Mora, Sergio |
Grupo/s de investigación o GITE: | Advanced deveLopment and empIrical research on Software (ALISoft) |
Centro, Departamento o Servicio: | Universidad de Alicante. Departamento de Lenguajes y Sistemas Informáticos |
Palabras clave: | Data mining | Data warehouse | Educational data mining | Academic development |
Área/s de conocimiento: | Lenguajes y Sistemas Informáticos |
Fecha de publicación: | 11-abr-2019 |
Editor: | Taylor & Francis |
Cita bibliográfica: | Australasian Journal of Engineering Education. 2019, 24(1): 4-13. doi:10.1080/22054952.2019.1601063 |
Resumen: | Nowadays, researchers analyse student data to predict the graduation rate by looking at the characteristics of students enrolled and to take corrective actions at an early stage or improve the admission process. Educational data mining (EDM) is an emerging field that can support the implementation of changes in the management of higher education institutions. EDM analyses educational data using the development and the application of data mining (DM) methods and algorithms to information stored in academic data repositories. The purpose of this paper is to review which methods and algorithms of DM can be used in the analysis of educational data to improve decision-making. Furthermore, it evaluates these algorithms using a dataset composed of student data in the computer science school of a private university. The core of the analysis is to discover trends and patterns of study in the graduation rate indicator. Finally, it compares these methods and algorithms and suggests which has the best precision in certain scenarios. Our analyses suggest that random trees had better precision but had limitations due to the difficulty of interpretation while the J48 algorithm had better possibilities of interpretation of results in the visualisation of the classification of data and only had slightly inferior performance. |
URI: | http://hdl.handle.net/10045/93995 |
ISSN: | 1325-4340 (Print) | 1324-5821 (Online) |
DOI: | 10.1080/22054952.2019.1601063 |
Idioma: | eng |
Tipo: | info:eu-repo/semantics/article |
Derechos: | © 2019 Engineers Australia |
Revisión científica: | si |
Versión del editor: | https://doi.org/10.1080/22054952.2019.1601063 |
Aparece en las colecciones: | INV - ALISoft - Artículos de Revistas |
Archivos en este ítem:
Archivo | Descripción | Tamaño | Formato | |
---|---|---|---|---|
2019_Moscoso-Zea_etal_AustralasianJEngEdu_final.pdf | Versión final (acceso restringido) | 1,42 MB | Adobe PDF | Abrir Solicitar una copia |
2019_Moscoso-Zea_etal_AustralasianJEngEdu_preprint.pdf | Preprint (acceso abierto) | 438,09 kB | Adobe PDF | Abrir Vista previa |
Todos los documentos en RUA están protegidos por derechos de autor. Algunos derechos reservados.