Modelos de aprendizaje automático para la predicción de la deserción escolar en contextos rurales
International Journal of Innovation and Pedagogy · 15.05.2026 · 3 (1) · 68-82
DOI: 10.64325/ijip.v3i1.56 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Innovation and Pedagogy, online erschienen am 15.05.2026. Verfasst von Oscar Iván Montiel Petro. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This research systematizes the academic evidence on the use of machine learning models to predict school dropout rates in basic and secondary education in rural contexts (2015-2025). A systematic review was conducted following PRISMA 2020 guidelines, and Scopus, Web of Science, SciELO, and Redalyc were consulted. After applying eligibility criteria focused on rurality and vulnerable populations, 25 articles were selected. The results show the predominance of ensemble models, especially Random Forest and XGBoost, due to their performance with scarce and unbalanced tabular data. A shift in the most influential predictors is also observed: structural and geographic variables—home-school distance, maternal educational level, and agricultural cycles—outperform prior academic performance. In parallel, the explanability of models (XAI) emerges as a key requirement for implementing Early Warning Systems that guide timely decisions. In summary, machine learning outperforms traditional statistical approaches for detecting dropout risk in rural areas. However, its effectiveness depends on integrating contextualized data and translating algorithmic results into pedagogical actions that are understandable for teachers, administrators, and local authorities. There are also reported concerns about the quality and availability of administrative data, as well as about ethics, privacy, and bias in rural communities.
Zitieren
APA
Petro, O. I. M. (2026). Modelos de aprendizaje automático para la predicción de la deserción escolar en contextos rurales. International Journal of Innovation and Pedagogy, 3(1), 68-82. https://doi.org/10.64325/ijip.v3i1.56
Vancouver
Petro OIM. Modelos de aprendizaje automático para la predicción de la deserción escolar en contextos rurales. International Journal of Innovation and Pedagogy. 2026;3(1):68-82. doi:10.64325/ijip.v3i1.56
BibTeX
@article{Petro2026,
author = {Petro, Oscar Iván Montiel},
title = {Modelos de aprendizaje automático para la predicción de la deserción escolar en contextos rurales},
journal = {International Journal of Innovation and Pedagogy},
year = {2026},
volume = {3},
number = {1},
pages = {68--82},
doi = {10.64325/ijip.v3i1.56}
}
EUVIMED