Revisiting the Dissimilarity Representation in the Context of Regression

In machine learning, a natural way to represent an instance is by using a feature vector. However, several studies have shown that this representation may not accurately characterize an object. For classification problems, the dissimilarity paradigm has been proposed as an alternative to the standar...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Vicente Garcia, J. Salvador Sanchez, Rafael Martinez-Pelaez, Luis C. Mendez-Gonzalez
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Data complexity dissimilarity representation linear models regression Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/bcf6a3c7b90c46798ec66106b046576c
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:bcf6a3c7b90c46798ec66106b046576c
record_format	dspace
spelling	oai:doaj.org-article:bcf6a3c7b90c46798ec66106b046576c2021-12-02T00:00:45ZRevisiting the Dissimilarity Representation in the Context of Regression2169-353610.1109/ACCESS.2021.3130127https://doaj.org/article/bcf6a3c7b90c46798ec66106b046576c2021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9624949/https://doaj.org/toc/2169-3536In machine learning, a natural way to represent an instance is by using a feature vector. However, several studies have shown that this representation may not accurately characterize an object. For classification problems, the dissimilarity paradigm has been proposed as an alternative to the standard feature-based approach. Encoding each object by pairwise dissimilarities has been demonstrated to improve the data quality because it mitigates some complexities such as class overlap, small disjuncts, and low-sample size. However, its suitability and performance when applied to regression problems have not been fully explored. This study redefines the dissimilarity representation for regression. To this end, we have carried out an extensive experimental evaluation on 34 datasets using two linear regression models. The results show that the dissimilarity approach decreases the error rates of both the traditional linear regression and the linear model with elastic net regularization, and it also reduces the complexity of most regression datasets.Vicente GarciaJ. Salvador SanchezRafael Martinez-PelaezLuis C. Mendez-GonzalezIEEEarticleData complexitydissimilarity representationlinear modelsregressionElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 157043-157051 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Data complexity dissimilarity representation linear models regression Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Data complexity dissimilarity representation linear models regression Electrical engineering. Electronics. Nuclear engineering TK1-9971 Vicente Garcia J. Salvador Sanchez Rafael Martinez-Pelaez Luis C. Mendez-Gonzalez Revisiting the Dissimilarity Representation in the Context of Regression
description	In machine learning, a natural way to represent an instance is by using a feature vector. However, several studies have shown that this representation may not accurately characterize an object. For classification problems, the dissimilarity paradigm has been proposed as an alternative to the standard feature-based approach. Encoding each object by pairwise dissimilarities has been demonstrated to improve the data quality because it mitigates some complexities such as class overlap, small disjuncts, and low-sample size. However, its suitability and performance when applied to regression problems have not been fully explored. This study redefines the dissimilarity representation for regression. To this end, we have carried out an extensive experimental evaluation on 34 datasets using two linear regression models. The results show that the dissimilarity approach decreases the error rates of both the traditional linear regression and the linear model with elastic net regularization, and it also reduces the complexity of most regression datasets.
format	article
author	Vicente Garcia J. Salvador Sanchez Rafael Martinez-Pelaez Luis C. Mendez-Gonzalez
author_facet	Vicente Garcia J. Salvador Sanchez Rafael Martinez-Pelaez Luis C. Mendez-Gonzalez
author_sort	Vicente Garcia
title	Revisiting the Dissimilarity Representation in the Context of Regression
title_short	Revisiting the Dissimilarity Representation in the Context of Regression
title_full	Revisiting the Dissimilarity Representation in the Context of Regression
title_fullStr	Revisiting the Dissimilarity Representation in the Context of Regression
title_full_unstemmed	Revisiting the Dissimilarity Representation in the Context of Regression
title_sort	revisiting the dissimilarity representation in the context of regression
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/bcf6a3c7b90c46798ec66106b046576c
work_keys_str_mv	AT vicentegarcia revisitingthedissimilarityrepresentationinthecontextofregression AT jsalvadorsanchez revisitingthedissimilarityrepresentationinthecontextofregression AT rafaelmartinezpelaez revisitingthedissimilarityrepresentationinthecontextofregression AT luiscmendezgonzalez revisitingthedissimilarityrepresentationinthecontextofregression
_version_	1718404011026022400

Revisiting the Dissimilarity Representation in the Context of Regression

Ejemplares similares