Predicción de riesgo de anemia para niños menores a cinco años en regiones del Perú usando datos del INEI
Average rating
Cast your vote
You can rate an item by clicking the amount of stars they wish to award to this item.
When enough users have cast their vote on this item, the average rating will also be shown.
Star rating
Your vote was cast
Thank you for your feedback
Thank you for your feedback
Authors
Silva Manayalle, RicardoEnriquez Ramirez, Stephan Andre
Quispe Manrique, Miguel Cleme
Teccsi Cerda, Diego Nelson
Advisors
VARGAS CIRILO, Hernán RogerIssue Date
2025-12-05Keywords
Anemia infantilMachine learning
Predicción de riesgo
Salud pública
ENDES
Factores socioeconómicos
Desigualdad regional
Childhood anemia
Risk prediction
Public health
Socioeconomic factors
Regional inequality
Metadata
Show full item recordOther Titles
Prediction of Anemia Risk Among Children Under Five in Peru Using INEI DataAbstract
La anemia infantil constituye un problema crítico de salud pública en el Perú, afectando al 33.7% de niños entre 6 y 59 meses, con marcadas disparidades entre zonas rurales (42.3%) y urbanas (30.2%). Esta investigación desarrolla un modelo predictivo de Machine Learning para identificar el riesgo de anemia en niños menores de 5 años, utilizando datos de la Encuesta Demográfica y de Salud Familiar (ENDES) del período 2022-2024. Se implementaron y evaluaron cinco algoritmos: Regresión Logística, Random Forest, XGBoost, LightGBM y CatBoost, aplicando la metodología CRISP-DM combinada con SCRUM. El análisis procesó 55,808 registros, generando 78 variables mediante ingeniería de características. Contrario a las expectativas iniciales, la Regresión Logística regularizada alcanzó el mejor desempeño con un AUC de 0.7208 y sensibilidad del 74.72%, identificando correctamente 7 de cada 10 casos de anemia. Los hallazgos revelan que los factores socioeconómicos, particularmente la pobreza multidimensional (importancia: 14.53%), predominan sobre los indicadores antropométricos tradicionales. La heterogeneidad regional emergió como determinante crítico, con ocho de las diez variables más predictivas correspondiendo a factores geográficos. El modelo permite la transición de un enfoque reactivo a uno proactivo en salud pública, facilitando la identificación temprana de poblaciones vulnerables y la focalización eficiente de recursos. Los resultados sugieren que combatir la anemia requiere intervenciones multisectoriales que aborden las desigualdades estructurales más allá de la suplementación nutricional tradicional.Childhood anemia constitutes a critical public health problem in Peru, affecting 33.7% of children between 6 and 59 months, with marked disparities between rural (42.3%) and urban (30.2%) areas. This research develops a Machine Learning predictive model to identify anemia risk in children under 5 years, using data from the Demographic and Family Health Survey (ENDES) for the period 2022-2024. Five algorithms were implemented and evaluated: Logistic Regression, Random Forest, XGBoost, LightGBM, and CatBoost, applying the CRISP-DM methodology combined with SCRUM. The analysis processed 55,808 records, generating 78 variables through feature engineering. Contrary to initial expectations, regularized Logistic Regression achieved the best performance with an AUC of 0.7208 and sensitivity of 74.72%, correctly identifying 7 out of 10 anemia cases. The findings reveal that socioeconomic factors, particularly multidimensional poverty (importance: 14.53%), predominate over traditional anthropometric indicators. Regional heterogeneity emerged as a critical determinant, with eight of the ten most predictive variables corresponding to geographic factors. The model enables the transition from a reactive to a proactive approach in public health, facilitating early identification of vulnerable populations and efficient resource allocation. Results suggest that combating anemia requires multisectoral interventions that address structural inequalities beyond traditional nutritional supplementation.
Type
http://purl.org/coar/resource_type/c_bdccRights
http://purl.org/coar/access_right/c_abf2Language
spaCollections
Except where otherwise noted, this item's license is described as http://purl.org/coar/access_right/c_abf2

