Un modelo Fay-Herriot espacial para la predicción del porcentaje de niños con anemia y riesgo de retraso del crecimiento en distritos no encuestados y en distritos con pocas observaciones disponibles
No hay miniatura disponible
Fecha
2023-02-24
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Pontificia Universidad Católica del Perú
DOI
Resumen
La presente tesis tiene como objetivo realizar predicciones del nivel de prevalencia de anemia y riesgo de retraso del crecimiento en niños menores de 5 años, en distritos encuestados con observaciones faltantes y distritos no encuestados con niveles de pobreza mayor a 55% del país, con datos de la Encuesta Demográfica de Salud Familiar (ENDES) para el año 2019 y el censo nacional del 2017. Para alcanzar este objetivo se hace uso del modelo Fay-Herriot y su variante espacial. Este modelo permite obtener estimaciones fiables en dominios en los cuales el diseño muestral no tiene un nivel de inferencia aceptable. El uso de variables auxiliares obtenidas del censo a nivel de distrito permite implementar el modelo Fay- Herriot, obteniendo predicciones con un error de estimación menor a las estimaciones muestrales. En el caso de la variante espacial del modelo Fay-Herriot, este incluye correlaciones entre los efectos aleatorios de áreas vecinas. Además, se estima el error cuadrático medio de las predicciones obtenidas mediante los métodos de bootstrap paramétrico y no paramétrico. Los resultados muestran que los distritos de los departamentos de Puno, Cusco y Huancavelica tienen indicadores de anemia en niños menores de 5 años que superan el 48.5% en tanto que los distritos de los departamentos de Puno, Huancavelica y Tacna tienen indicadores de riesgo de retraso del crecimiento en niños menores de 5 años por encima del 45.5 %.
The objective of this thesis is to make predictions of the level of prevalence of anemia and risk of stunted growth in children under 5 years of age, in surveyed districts with missing observations and non-surveyed districts with poverty levels greater than 55% of the country, with data from the Family Health Demo- graphic Survey (ENDES) for the year 2019 and the national census of 2017. To achieve this objective, the Fay-Herriot model and its spatial variant are used. This model makes it possible to obtain reliable estimates in domains in which the sample design does not have an acceptable level of inference. The use of auxiliary variables obtained from the census at the district level makes it possible to implement the Fay-Herriot model, obtaining predictions with an estimation error smaller than the sample estimates. In the case of the spatial variant of the Fay-Herriot model, it includes correlations between the random effects of neighboring areas. In addition, the mean square error of the predictions obtained using the parametric and non-parametric bootstrap methods is estimated. The results show that the districts of the departments of Puno, Cusco and Huancavelica have indicators of anemia in children under 5 years of age that exceed 48.5 %, while the districts of the departments of Puno, Huancavelica and Tacna have indicators of risk of stunted growth in children under 5 years of age above 45.5 %.
The objective of this thesis is to make predictions of the level of prevalence of anemia and risk of stunted growth in children under 5 years of age, in surveyed districts with missing observations and non-surveyed districts with poverty levels greater than 55% of the country, with data from the Family Health Demo- graphic Survey (ENDES) for the year 2019 and the national census of 2017. To achieve this objective, the Fay-Herriot model and its spatial variant are used. This model makes it possible to obtain reliable estimates in domains in which the sample design does not have an acceptable level of inference. The use of auxiliary variables obtained from the census at the district level makes it possible to implement the Fay-Herriot model, obtaining predictions with an estimation error smaller than the sample estimates. In the case of the spatial variant of the Fay-Herriot model, it includes correlations between the random effects of neighboring areas. In addition, the mean square error of the predictions obtained using the parametric and non-parametric bootstrap methods is estimated. The results show that the districts of the departments of Puno, Cusco and Huancavelica have indicators of anemia in children under 5 years of age that exceed 48.5 %, while the districts of the departments of Puno, Huancavelica and Tacna have indicators of risk of stunted growth in children under 5 years of age above 45.5 %.
Descripción
Palabras clave
Variables (Estadística), Inferencia estadística--Anemia--Perú, Encuestas de hogares--Perú
Citación
Colecciones
item.page.endorsement
item.page.review
item.page.supplemented
item.page.referenced
Licencia Creative Commons
Excepto se indique lo contrario, la licencia de este artículo se describe como info:eu-repo/semantics/openAccess