Machine learning model for predicting school dropout through household surveys

• Autores: Jacha Rojas, Johnny P., Yataco Cañari, Walter, Ospina Galindez, Johann A.
• Formato: artículo
• Fecha de Publicación: 2024
• Institución: Universidad Nacional Hermilio Valdizan
• Repositorio: Revistas - Universidad Nacional Hermilio Valdizán
• Lenguaje: español
• OAI Identifier: oai:revistas.unheval.edu.pe:article/2308
• Enlace del recurso: http://revistas.unheval.edu.pe/index.php/rifce/article/view/2308
• Nivel de acceso: acceso abierto
• Materia: Estrategias; didácticas; interactivas del docente; strategies; didactic; teacher interactive

The persistent issue of student dropout negatively impacts the educational sector and society at large. This study presents a machine learning model that leverages data from the National Household Survey to predict student dropout in Peru, integrating a wide range of socio-demographic variables. The research fills a gap in existing literature by providing a model that incorporates socio-demographic variables, an area not fully explored in previous studies. The predictive model aims to identify factors associated with student dropout, aiding educational stakeholders in implementing effective interventions. The findings underscore the model's potential to enhance educational outcomes by enabling early identification of at-risk students, thereby facilitating targeted support. This work contributes to refining predictive models of university dropout rates and sug- gests the use of ensemble methods to improve the accu- racy of single-model predictions. Future research could further explore computational methodologies like deep learning and hybrid models to predict dropout rates and their comparison with this study's outcomes, considering additional influential factors not covered in this research.