Prediction of academic performance using data mining in first year students of peruvian university

• Autores: Yamao, Eiriku, Celi Saavedra, Luis, Campos Pérez, Rosalvina, Huancas Hurtado, Valery de Jesús
• Formato: artículo
• Fecha de Publicación: 2018
• Institución: Universidad de San Martín de Porres
• Repositorio: Revistas - Universidad de San Martín de Porres
• Lenguaje: español
• OAI Identifier: oai:revistas.usmp.edu.pe:article/1371
• Enlace del recurso: https://portalrevistas.aulavirtualusmp.pe/index.php/rc/article/view/1371
• Nivel de acceso: acceso abierto
• Materia: Academic Performance; prediction; Educational Data Mining; EDM; Higher Education

Academic performance is a subject that has been studied for a long time. First year students in universities are the most vulnerable to face performance problems, resulting in possible desertion. Data mining in education applies data mining techniques in the information generated in the education sector. The present research consists of making the prediction of the academic performance of the students who entered the Professional School of Computer and Systems Engineering of the University of San Martín de Porres in the first cycle using data mining. Data were extracted from 1304 entrants who were classified using three factors: social, economic and academic, and predictions were made using three techniques: linear regression, decision tree and support vector machines, having the best result of 82.87% obtained using the decision tree. Out of the different factors, those that most influenced the academic performance were the following: admission exam grade, gender, age, income and distance from home to the study center. Using data mining it was possible to elaborate predictions of the academic performance of the students, which allowed the detection of students who could encounter issues in their studies during the first semester.