Use this link to cite:
http://hdl.handle.net/2183/39831 Análisis comparativo de técnicas de aprendizaje estadístico en R: Aplicación a modelos de regresión
Loading...
Identifiers
Publication date
Authors
Rama Martínez, Iago
Advisors
Other responsabilities
Universidade da Coruña. Facultade de Informática
Journal Title
Bibliographic citation
Type of academic work
Academic degree
Abstract
[Resumen]: El objetivo de este Trabajo de Fin de Grado es la realización de un análisis comparativo de dos metapaquetes (Caret y Tidymodels) del software R destinados al modelado de modelos de aprendizaje automático. Aunque estos metapaquetes no estén orientados en las tareas de resumen y visualización de datos, se evaluará su capacidad para realizar Analisis Exploratorio de Datos sobre un conjunto de datos. Se buscará proporcionar información de que metapaquete aporta más recursos para realizar esta tarea concreta. Se compararán los errores de entrenamiento y test devueltos por los modelos generados con cada metapaquete, buscando decidir cuál de ellos genera modelos más precisos. Finalmente, se medirán los tiempos de ejecución a la hora de entrenar los modelos de regresión para así determinar el rendimiento de dichos modelos, empleando tanto computación secuencial como computación paralela. En conclusión, con este estudio se pretende determinar las ventajas e inconvenientes del uso de cada uno de los metapaquetes a la hora de trabajar con modelos de regresión.
[Abstract]: The objective of this Final Degree Project is to carry out a comparative analysis of two metapackages (Caret and Tidymodels) of the R software intended for modeling machine learning models. Although these metapackages are not oriented towards data summary and visualization tasks, their capacity to perform Exploratory Data Analysis on a set of data will be evaluated. The aim is to provide information on which metapackage provides more resources to carry out this specific task. The training and test errors returned by the models generated with each metapackage will be compared, seeking to decide which of them generates more precise models. Finally, the execution times when training the regression models will be measured in order to determine the performance of said models, using both sequential computing and parallel computing. In conclusion, this study aims to determine the advantages and disadvantages of using each of the metapackages when working with regression models.
[Abstract]: The objective of this Final Degree Project is to carry out a comparative analysis of two metapackages (Caret and Tidymodels) of the R software intended for modeling machine learning models. Although these metapackages are not oriented towards data summary and visualization tasks, their capacity to perform Exploratory Data Analysis on a set of data will be evaluated. The aim is to provide information on which metapackage provides more resources to carry out this specific task. The training and test errors returned by the models generated with each metapackage will be compared, seeking to decide which of them generates more precise models. Finally, the execution times when training the regression models will be measured in order to determine the performance of said models, using both sequential computing and parallel computing. In conclusion, this study aims to determine the advantages and disadvantages of using each of the metapackages when working with regression models.
Description
Editor version
Rights
Atribución-NoComercial 3.0 España







