Estimación de conjuntos de nivel para el COVID-2019
Loading...
Identifiers
Publication date
Authors
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
[ES] La pandemia producida por el COVID-19 está causando grandes estragos económicos, sociales y sanitarios a lo largo del mundo. En vista de la falta de recursos, encontrar aquellas zonas más perjudicadas permitiría enviar ayudas localizadas y frenar al virus más rápida y eficazmente. Este será el objetivo de este trabajo en el que se reconstruirán concretamente los focos de contagios por coronavirus en Estados Unidos mediante la estimación no paramétrica de conjuntos de nivel. En particular, examinaremos aquellos conjuntos que superen cierto umbral de probabilidad denominados regiones de alta densidad. Reconstruirlos supone estimar la función densidad, un problema bien conocido en la Estadística No Paramétrica. Para ello, el estimador más utilizado es el de tipo núcleo en cuya definición aparece una matriz ventana H que debe especificar el usuario y que influye notablemente en el resultado. Se revisará la teoría de cuatro métodos diferentes de selección de H: método de validación cruzada por mínimos cuadrados, método plug-in de Duong y Hazelton, método de validación cruzada sesgada y método de validación cruzada suavizado. Posteriormente, se aplicarán tres de estos métodos a nuestros datos y se compararán las soluciones obtenidas.
[EN] The COVID-19 pandemic is wreaking great economic, social and health havoc across the world. Due to the lack of resources, targeting the worst-hit areas would allow us to send localized aid and stop the virus more quickly and effectively. This will be the objective of this work in which coronavirus outbreaks in the United States will be reconstruct using nonparametric estimation of level sets. Specifically, sets that exceed a certain probability threshold called high density regions will be examined. Recomposing them involves estimating the density function, a well-known problem in Nonparametric Statistics. For this purpose the most commonly used estimator is the kernel type, whose definition includes a window matrix that must be specified by the user and which has a significant influence on the result. The theory of four different methods of selecting H will be reviewed: least squares cross validation method, Duong and Hazelton plug-in method, biased cross validation method and smoothed cross validation method. Three of these will then be applied to our data and the solutions obtained will be compared.
[EN] The COVID-19 pandemic is wreaking great economic, social and health havoc across the world. Due to the lack of resources, targeting the worst-hit areas would allow us to send localized aid and stop the virus more quickly and effectively. This will be the objective of this work in which coronavirus outbreaks in the United States will be reconstruct using nonparametric estimation of level sets. Specifically, sets that exceed a certain probability threshold called high density regions will be examined. Recomposing them involves estimating the density function, a well-known problem in Nonparametric Statistics. For this purpose the most commonly used estimator is the kernel type, whose definition includes a window matrix that must be specified by the user and which has a significant influence on the result. The theory of four different methods of selecting H will be reviewed: least squares cross validation method, Duong and Hazelton plug-in method, biased cross validation method and smoothed cross validation method. Three of these will then be applied to our data and the solutions obtained will be compared.
Description
Traballo Fin de Grao en Matemáticas. Curso 2020-2021
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







