Por favor, use este identificador para citar o enlazar este ítem:
http://cimat.repositorioinstitucional.mx/jspui/handle/1008/968| SMALL AREA ESTIMATION BASED ON A TWO-FOLD NESTED ERROR LOGNORMAL MODEL | |
| Georges Bucyibaruta | |
| Acceso Abierto | |
| Atribución-NoComercial | |
| empirical bayes predictor | |
| The demand for reliable small area estimates derived from survey data has increased greatly in recent years due to, among other factors, their growing use in formulating policies and programs, allocation of government funds, regional planning, small area business decisions and other applications. Following the denition given by Rao (2003), the term small area or small domain refers to a subpopulation for which the domain-specic sample is not large enough to produce direct estimates with reliable precision. This subpopulation can be a small geographical area (county, state, district, etc.), a demographic group within a geographical region (specic sex-age group, etc.) or any subdivision of the population. Most surveys provide little information on individual small areas since they are generally designed to produce accurate estimates at a higher level of aggregation. Small area estimation methods are well suited for settings that involve many domains, with small (or no) samples from individual domains. In this setting, traditional designbased direct survey estimates based only on samples from individual small areas are not reliable. In order to improve on the traditional estimates based on individual area sample, one may "borrow strength" from neighboring or related small areas, or other correlated dependent variables and relevant covariate information available from other sources, such as administrative records, to produce accurate small area estimates (Molina and Rao, 2015). Most of the methods proposed in literature, to estimate small-area quantities, assume that the variable of interest follows a linear model and the linking covariates are available at population element (or observational unit) level or area level (Pereira and Coelho, 2012; Marhuenda et al., 2013; Pfeermann, D., 2013; Petrucci and Pratesi, 2014; Berg and Chandra, 2014; Rao and Molina , 2015). This assumption is not common, as it is plausible that some variables of interest in various surveys can be skewed distributed (Molina and Rao, 2010; Karlberg, 2014). Besides, it is not always easy to link the covariates obtained from other sources (censuses and/or other surveys) to those associated with the characteristic of interest (Datta and Ghosh, 1991). In this thesis, we consider small area estimation (SAE) techniques focusing primarily on estimating and predicting skewed (lognormal) distributed data. A brief review on the theory of Linear Mixed Models is given. Estimation of a small area | |
| 02-12-2018 | |
| Tesis de doctorado | |
| Inglés | |
| PROBABILIDAD | |
| Versión aceptada | |
| acceptedVersion - Versión aceptada | |
| Aparece en las colecciones: | Tesis del CIMAT |
Cargar archivos:
| Fichero | Descripción | Tamaño | Formato | |
|---|---|---|---|---|
| TE 692.pdf | 642.42 kB | Adobe PDF | Visualizar/Abrir |