Asociación entre variables cualitativas: ejemplo práctico de análisis de correspondencias
Loading...
Identifiers
Publication date
Authors
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
[ES] El análisis de correspondencias es una técnica de la estadística descriptiva aplicada a
tablas de contingencia cuyo objetivo es la visualización de una nube de puntos multidimensional
asociada a unas ciertas variables en un espacio de menor dimensión. Se generan
mapas que plasman los puntos intentando respetar al máximo sus posiciones en el espacio
original perdiendo el mínimo de información, lo que nos permite elaborar un análisis exploratorio
de ellos rápido y eficaz además de poder establecer asociaciones entre las variables.
Las proyecciones habitualmente son sobre espacios de dimensión 2 por nuestra familiaridad
con ellos.
Se tratará el contraste de la homogeneidad de la tabla, se introducirá una medida
más adecuada para esta técnica (la distancia ji-cuadrado) y se expresará la varianza de la
tabla en función de estos conceptos. Esta nueva medida de la varianza es lo que queremos
conservar al realizar la proyección de la tabla. Se introducirán diversos tipos de análisis
de correspondencias y técnicas para su implementación, así como realizar inferencia y
contrastes sobre los datos.
Todo el trabajo se acompaña de varios ejemplos para facilitar la comprensión de qué
se está haciendo. Por último, llevaremos a la práctica todo lo introducido a lo largo del
trabajo en un estudio médico real con las pretensiones de resumir los datos, establecer un
análisis descriptivo y establecer relaciones entre varias enfermedades comunes y la calidad
y forma de vida de 820 pacientes.
[EN] Correspondence analysis is a technique of descriptive statistics applied to contingency tables whose objective is the visualization of a multidimensional point cloud associated to certain variables in a space of smaller dimension. Maps are generated trying to respect as much as possible their positions in the original space, losing the minimum of information, which allows us to elaborate an exploratory analysis of them quickly and eficiently as well as being able to establish associations between the variables. The projections are usually on spaces of dimension 2 because of our familiarity with them. The contrast of the homogeneity of the table will be treated, a more appropriate measure will be introduced for this technique (the chi-square distance) and the variance of the table will be expressed in terms of these concepts. This new measure of variance is what we want to keep when projecting the table. Various types of correspondence analysis and techniques for its implementation will be introduced, as well as inference and contrasts on the data. All the work is accompanied by several examples to facilitate the understanding of what is being done. Finally, we will put into practice everything introduced during the work in a real medical study with the aim of summarizing the data, establishing a descriptive analysis and establishing relationships between several common diseases and the quality and way of life of 820 patients.
[EN] Correspondence analysis is a technique of descriptive statistics applied to contingency tables whose objective is the visualization of a multidimensional point cloud associated to certain variables in a space of smaller dimension. Maps are generated trying to respect as much as possible their positions in the original space, losing the minimum of information, which allows us to elaborate an exploratory analysis of them quickly and eficiently as well as being able to establish associations between the variables. The projections are usually on spaces of dimension 2 because of our familiarity with them. The contrast of the homogeneity of the table will be treated, a more appropriate measure will be introduced for this technique (the chi-square distance) and the variance of the table will be expressed in terms of these concepts. This new measure of variance is what we want to keep when projecting the table. Various types of correspondence analysis and techniques for its implementation will be introduced, as well as inference and contrasts on the data. All the work is accompanied by several examples to facilitate the understanding of what is being done. Finally, we will put into practice everything introduced during the work in a real medical study with the aim of summarizing the data, establishing a descriptive analysis and establishing relationships between several common diseases and the quality and way of life of 820 patients.
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



