Support Vector Components Analysis

Michiel van der Ree; Johannes Roerdink; Christophe Phillips; Gaetan Garraux; Eric Salmon; Marco Wiering

Support Vector Components Analysis

Michiel van der Ree, Johannes Roerdink, Christophe Phillips, Gaetan Garraux, Eric Salmon, Marco Wiering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

119 Downloads (Pure)

Abstract

In this paper we propose a novel method for learning a distance metric in the process of training Support Vector Machines (SVMs) with the radial basis function kernel. A transformation matrix is adapted in such a way that the SVM dual objective of a classification problem is optimized. By using a wide transformation matrix the method can effectively be used as a means of supervised dimensionality reduction. We compare our method with other algorithms on a toy dataset and on PET-scans of patients with various Parkinsonisms, finding that our method either outperforms or performs on par with the other algorithms.

Original language	English
Title of host publication	European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Subtitle of host publication	ESANN
Publisher	ESANN
Number of pages	6
ISBN (Print)	978-287587039-1
Publication status	Published - 26-Apr-2017
Event	ESANN 2017 - 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning - Bruges, Belgium Duration: 26-Apr-2017 → 28-Nov-2017

Conference

Conference	ESANN 2017 - 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Country/Territory	Belgium
City	Bruges
Period	26/04/2017 → 28/11/2017

Keywords

Support vector machines
Dimensionality reduction
Machine Learning

Access to Document

Support Vector Components AnalysisFinal author's version, 290 KB

Handle.net

Persistent link

Cite this

@inproceedings{0046e6e535014280a89550675656680f,

title = "Support Vector Components Analysis",

abstract = "In this paper we propose a novel method for learning a distance metric in the process of training Support Vector Machines (SVMs) with the radial basis function kernel. A transformation matrix is adapted in such a way that the SVM dual objective of a classification problem is optimized. By using a wide transformation matrix the method can effectively be used as a means of supervised dimensionality reduction. We compare our method with other algorithms on a toy dataset and on PET-scans of patients with various Parkinsonisms, finding that our method either outperforms or performs on par with the other algorithms.",

keywords = "Support vector machines, Dimensionality reduction, Machine Learning",

author = "{van der Ree}, Michiel and Johannes Roerdink and Christophe Phillips and Gaetan Garraux and Eric Salmon and Marco Wiering",

year = "2017",

month = apr,

day = "26",

language = "English",

isbn = "978-287587039-1",

booktitle = "European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning",

publisher = "ESANN",

note = "ESANN 2017 - 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning ; Conference date: 26-04-2017 Through 28-11-2017",

}

van der Ree, M, Roerdink, J, Phillips, C, Garraux, G, Salmon, E & Wiering, M 2017, Support Vector Components Analysis. in European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning: ESANN. ESANN, ESANN 2017 - 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, Bruges, Belgium, 26/04/2017.

TY - GEN

T1 - Support Vector Components Analysis

AU - van der Ree, Michiel

AU - Roerdink, Johannes

AU - Phillips, Christophe

AU - Garraux, Gaetan

AU - Salmon, Eric

AU - Wiering, Marco

PY - 2017/4/26

Y1 - 2017/4/26

N2 - In this paper we propose a novel method for learning a distance metric in the process of training Support Vector Machines (SVMs) with the radial basis function kernel. A transformation matrix is adapted in such a way that the SVM dual objective of a classification problem is optimized. By using a wide transformation matrix the method can effectively be used as a means of supervised dimensionality reduction. We compare our method with other algorithms on a toy dataset and on PET-scans of patients with various Parkinsonisms, finding that our method either outperforms or performs on par with the other algorithms.

AB - In this paper we propose a novel method for learning a distance metric in the process of training Support Vector Machines (SVMs) with the radial basis function kernel. A transformation matrix is adapted in such a way that the SVM dual objective of a classification problem is optimized. By using a wide transformation matrix the method can effectively be used as a means of supervised dimensionality reduction. We compare our method with other algorithms on a toy dataset and on PET-scans of patients with various Parkinsonisms, finding that our method either outperforms or performs on par with the other algorithms.

KW - Support vector machines

KW - Dimensionality reduction

KW - Machine Learning

M3 - Conference contribution

SN - 978-287587039-1

BT - European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning

PB - ESANN

T2 - ESANN 2017 - 25th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning

Y2 - 26 April 2017 through 28 November 2017

ER -

Support Vector Components Analysis

Abstract

Conference

Keywords

Access to Document

Handle.net

Fingerprint

Cite this