Inferencia estadística con datos sesgados por longitud
Loading...
Identifiers
Publication date
Authors
Advisors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
[ES] En este trabajo se hace un recorrido a lo largo de varias situaciones que nos llevan a la
aparición de sesgo por longitud en una muestra. Presentaremos el modelo de sesgo por longitud,
y a partir de él veremos que el estimador adecuado de la media poblacional no es la
media aritmética simple, sino que es la media armónica. Mediante un estudio de simulación
analizaremos las propiedades de la media armónica y comprobaremos que la media aritmé-
tica no es un buen estimador de la media poblacional, en el caso de tener muestras sesgadas.
También trabajaremos los métodos de estimación paramétrica más habituales (mínimos
cuadrados ponderados) adaptados al sesgo por longitud. Y empleando el programa ℜ
generaremos muestras sesgadas de un modelo de regresión y estimaremos los coeficientes
con y sin ponderación, para hacer una comparación de los sesgos y desviaciones típicas
aproximadas.
[EN] Several situations that exhibit biased sampling are presented in this paper. The length bias model is introduced and we will see that the appropriate estimator of the population mean is not the simple arithmetic mean, but is the harmonic mean. Through a simulation study, we will analyze the properties of the harmonic mean and we will verify that the arithmetic mean is not a good estimator of the population mean in the case of biased samples. We will also work with the most usual parametric estimation methods (weighted least squares) adapted to the length bias. And using the ℜ program, we will generate biased samples of a regression model, and we will estimate the coefficients with and without weighting, to make a comparison of approximate biases and standard deviations.
[EN] Several situations that exhibit biased sampling are presented in this paper. The length bias model is introduced and we will see that the appropriate estimator of the population mean is not the simple arithmetic mean, but is the harmonic mean. Through a simulation study, we will analyze the properties of the harmonic mean and we will verify that the arithmetic mean is not a good estimator of the population mean in the case of biased samples. We will also work with the most usual parametric estimation methods (weighted least squares) adapted to the length bias. And using the ℜ program, we will generate biased samples of a regression model, and we will estimate the coefficients with and without weighting, to make a comparison of approximate biases and standard deviations.
Description
Traballo Fin de Grao en Matemáticas. Curso 2018-2019
Keywords
Bibliographic citation
Relation
Has part
Has version
Is based on
Is part of
Is referenced by
Is version of
Requires
Sponsors
Rights
Atribución-NoComercial-CompartirIgual 4.0 Internacional



