Clustering Categories in Support Vector Machines

Emilio Carrizosa, Amaya Nogales-Gómez, Dolores Romero Morales

Research output: Working paperResearchpeer-review

Abstract

Support Vector Machines (SVM) is the state-of-the-art in Supervised Classification. In this paper the Cluster Support Vector Machines (CLSVM) methodology is proposed with the aim to reduce the complexity of the SVM classifier in the presence of categorical features. The CLSVM methodology lets categories cluster around their peers and builds an SVM classifier using the clustered dataset. Four strategies for building the CLSVM classifier are presented based on solving: the original SVM formulation, a Quadratically Constrained Quadratic Programming formulation, and a Mixed Integer Quadratic Programming formulation as well as its continuous relaxation. The computational study illustrates the performance of the CLSVM classifier using two clusters. In the tested datasets our methodology achieves comparable accuracy to that of the SVM with original data but with a dramatic decrease in complexity.
Original languageEnglish
Place of Publicationwww
PublisherMathematical Optimization Society
Number of pages20
Publication statusPublished - 2014
Externally publishedYes
SeriesOptimization Online
Number4403
Volume06

Cite this

Carrizosa, E., Nogales-Gómez, A., & Romero Morales, D. (2014). Clustering Categories in Support Vector Machines. www: Mathematical Optimization Society. Optimization Online, No. 4403, Vol.. 06
Carrizosa, Emilio ; Nogales-Gómez, Amaya ; Romero Morales, Dolores . / Clustering Categories in Support Vector Machines. www : Mathematical Optimization Society, 2014. (Optimization Online; No. 4403, Vol. 06).
@techreport{e863502c734a40958328dfe1ad784b05,
title = "Clustering Categories in Support Vector Machines",
abstract = "Support Vector Machines (SVM) is the state-of-the-art in Supervised Classification. In this paper the Cluster Support Vector Machines (CLSVM) methodology is proposed with the aim to reduce the complexity of the SVM classifier in the presence of categorical features. The CLSVM methodology lets categories cluster around their peers and builds an SVM classifier using the clustered dataset. Four strategies for building the CLSVM classifier are presented based on solving: the original SVM formulation, a Quadratically Constrained Quadratic Programming formulation, and a Mixed Integer Quadratic Programming formulation as well as its continuous relaxation. The computational study illustrates the performance of the CLSVM classifier using two clusters. In the tested datasets our methodology achieves comparable accuracy to that of the SVM with original data but with a dramatic decrease in complexity.",
keywords = "Support vector machines, Categorical features, Classifier complexity, Clustering, Quadratically constrained programming, 0-1 programming",
author = "Emilio Carrizosa and Amaya Nogales-G{\'o}mez and {Romero Morales}, Dolores",
year = "2014",
language = "English",
publisher = "Mathematical Optimization Society",
address = "United States",
type = "WorkingPaper",
institution = "Mathematical Optimization Society",

}

Carrizosa, E, Nogales-Gómez, A & Romero Morales, D 2014 'Clustering Categories in Support Vector Machines' Mathematical Optimization Society, www.

Clustering Categories in Support Vector Machines. / Carrizosa, Emilio; Nogales-Gómez, Amaya; Romero Morales, Dolores .

www : Mathematical Optimization Society, 2014.

Research output: Working paperResearchpeer-review

TY - UNPB

T1 - Clustering Categories in Support Vector Machines

AU - Carrizosa, Emilio

AU - Nogales-Gómez, Amaya

AU - Romero Morales, Dolores

PY - 2014

Y1 - 2014

N2 - Support Vector Machines (SVM) is the state-of-the-art in Supervised Classification. In this paper the Cluster Support Vector Machines (CLSVM) methodology is proposed with the aim to reduce the complexity of the SVM classifier in the presence of categorical features. The CLSVM methodology lets categories cluster around their peers and builds an SVM classifier using the clustered dataset. Four strategies for building the CLSVM classifier are presented based on solving: the original SVM formulation, a Quadratically Constrained Quadratic Programming formulation, and a Mixed Integer Quadratic Programming formulation as well as its continuous relaxation. The computational study illustrates the performance of the CLSVM classifier using two clusters. In the tested datasets our methodology achieves comparable accuracy to that of the SVM with original data but with a dramatic decrease in complexity.

AB - Support Vector Machines (SVM) is the state-of-the-art in Supervised Classification. In this paper the Cluster Support Vector Machines (CLSVM) methodology is proposed with the aim to reduce the complexity of the SVM classifier in the presence of categorical features. The CLSVM methodology lets categories cluster around their peers and builds an SVM classifier using the clustered dataset. Four strategies for building the CLSVM classifier are presented based on solving: the original SVM formulation, a Quadratically Constrained Quadratic Programming formulation, and a Mixed Integer Quadratic Programming formulation as well as its continuous relaxation. The computational study illustrates the performance of the CLSVM classifier using two clusters. In the tested datasets our methodology achieves comparable accuracy to that of the SVM with original data but with a dramatic decrease in complexity.

KW - Support vector machines

KW - Categorical features

KW - Classifier complexity

KW - Clustering

KW - Quadratically constrained programming

KW - 0-1 programming

M3 - Working paper

BT - Clustering Categories in Support Vector Machines

PB - Mathematical Optimization Society

CY - www

ER -

Carrizosa E, Nogales-Gómez A, Romero Morales D. Clustering Categories in Support Vector Machines. www: Mathematical Optimization Society. 2014.