Estabilización de la varianza
Loading...
Identifiers
Publication date
Authors
Advisors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
[ES] En el contexto de la construcción de intervalos de confianza para un parámetro desconocido,
es frecuente enfrentarse al problema de no poder construirlo debido a que el error
típico del estimador depende de dicho parámetro. Para solventar esta dificultad se estudiar
án dos métodos: el que denominaremos usual, que se basa en la estimación del error típico
y el de la transformación estabilizadora de la varianza, que es una aplicación del método
delta.
Se hará un estudio de simulación para la comparación de los intervalos de confianza
obtenidos por ambos métodos, en los que analizaremos las coberturas obtenidas para tres
tamaños muestrales, los tres niveles de confianza usuales y distintos valores de los verdaderos
parámetros. Además se añadirán distintas representaciones gráficas que muestran que
con el método de la transformación estabilizadora de la varianza se suele obtener una mejor
aproximación por la distribución normal y se suele mejorar la simetría de la distribución.
[EN] In the context of the construction of confidence intervals for an unknown parameter, it is frequent to face the problem of not being able to constuct said confidence interval due to the fact that standard error of the estimator depends on that parameter. In order to solve such a difficulty, two methods will be studied: the one we will refer to as usual, which is based on the estimations of the standard error, and that of the variance stabilizing transformation, which is an application of the delta method. A simulation study will be conducted for the comparison of the confidence intervals obtained with both methods, in which we will analize the coverages obtained for three sample sizes, the three usual levels of confidence and different values of the true parameters. Moreover, different graphic representations will be added that show that by using the variance-stabilizing transformation, a better approximation for the normal distribution is obtained. Additionally, the symetry of the distribution is often improved.
[EN] In the context of the construction of confidence intervals for an unknown parameter, it is frequent to face the problem of not being able to constuct said confidence interval due to the fact that standard error of the estimator depends on that parameter. In order to solve such a difficulty, two methods will be studied: the one we will refer to as usual, which is based on the estimations of the standard error, and that of the variance stabilizing transformation, which is an application of the delta method. A simulation study will be conducted for the comparison of the confidence intervals obtained with both methods, in which we will analize the coverages obtained for three sample sizes, the three usual levels of confidence and different values of the true parameters. Moreover, different graphic representations will be added that show that by using the variance-stabilizing transformation, a better approximation for the normal distribution is obtained. Additionally, the symetry of the distribution is often improved.
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



