Predición electoral mediante promedios de enquisas
Loading...
Identifiers
Publication date
Authors
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
[GL] Ante un escenario electoral xorde o interese de anticipar quen gañará e como será o
reparto de votos resultante. Nas semanas e meses previos véñense facendo sondaxes por
parte de diferentes casas de enquisas. Neste contexto, aparece a idea de agregar enquisas,
isto é, considerar un conxunto amplo de enquisas feitas ata ese momento e usalas para facer
un promedio que prediga o resultado o día das eleccións. Este será un promedio local que
priorice as enquisas máis próximas ó día do que se quere facer a predición sobre aquelas
que estean máis afastadas deste día.
Para este promedio local empregaremos un modelo estadístico de regresión non paramétrica
que recibe o nome de regresión lineal local. Así, durante os tres primeiros capítulos
farase un desenvolvemento teórico detallado deste modelo cubrindo os seus aspectos fundamentais,
como poden ser o cálculo dos estimadores ou a elección do parámetro ventana
que será o que determine como de local será o modelo. Finalmente, aplicarase todo este
desenvolvemento teórico ó caso real das eleccións presidenciais estadounidenses do 3 de
novembro do 2020.
[EN] In the presence of an electoral scenario, an interest arises in anticipating who will win and how the resulting vote distribution will be. In the previous weeks and months, polls have been carried out by different survey companies. In this context, the idea of aggregating polls appears, that is, considering a broad set of polls completed up to that point and using them to obtain an average that predicts the result on election day. This will be a local average that prioritizes the polls closest to the day for which you want to make the prediction over those that are further away from this day. For this local average, we will use a nonparametric regression statistical model that is called local linear regression. Thus, during the first three chapters a detailed theoretical development of this model will be made, covering its fundamental aspects, such as the calculation of the estimators or the choice of the window parameter that will determine how local the model will be. Finally, all of this theoretical development will be applied to the real case of the US presidential elections of the third of November 2020.
[EN] In the presence of an electoral scenario, an interest arises in anticipating who will win and how the resulting vote distribution will be. In the previous weeks and months, polls have been carried out by different survey companies. In this context, the idea of aggregating polls appears, that is, considering a broad set of polls completed up to that point and using them to obtain an average that predicts the result on election day. This will be a local average that prioritizes the polls closest to the day for which you want to make the prediction over those that are further away from this day. For this local average, we will use a nonparametric regression statistical model that is called local linear regression. Thus, during the first three chapters a detailed theoretical development of this model will be made, covering its fundamental aspects, such as the calculation of the estimators or the choice of the window parameter that will determine how local the model will be. Finally, all of this theoretical development will be applied to the real case of the US presidential elections of the third of November 2020.
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







