Repository logo
 
Publication

Improving Visualization, Scalability and Performance of Multiclass Problems with SVM Manifold Learning

2009-04Book part Open access
Simple item page
datacite.subject.fosCiências Naturais::Ciências da Computação e da Informação
datacite.subject.fosCiências Naturais::Matemáticas
dc.contributor.authorSilva, Catarina
dc.contributor.authorBernardete Ribeiro
dc.date.accessioned2025-05-30T12:06:33Z
dc.date.available2025-05-30T12:06:33Z
dc.date.issued2009-04
dc.description9th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 200923 April 2009 through 25 April 2009 - Code 83287
dc.description.abstractWe propose a learning framework to address multiclass challenges, namely visualization, scalability and performance. We focus on supervised problems by presenting an approach that uses prior information about training labels, manifold learning and support vector machines (SVMs). We employ manifold learning as a feature reduction step, nonlinearly embedding data in a low dimensional space using Isomap (Isometric Mapping), enhancing geometric characteristics and preserving the geodesic distance within the manifold. Structured SVMs are used in a multiclass setting with benefits for final multiclass classification in this reduced space. Results on a text classification toy example and on ISOLET, an isolated letter speech recognition problem, demonstrate the remarkable visualization capabilities of the method for multiclass problems in the severely reduced space, whilst improving SVMs baseline performance.eng
dc.identifier.citationSilva, C., Ribeiro, B. (2009). Improving Visualization, Scalability and Performance of Multiclass Problems with SVM Manifold Learning. In: Kolehmainen, M., Toivanen, P., Beliczynski, B. (eds) Adaptive and Natural Computing Algorithms. ICANNGA 2009. Lecture Notes in Computer Science, vol 5495. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-04921-7_38.
dc.identifier.doi10.1007/978-3-642-04921-7_38
dc.identifier.eissn1611-3349
dc.identifier.isbn9783642049200
dc.identifier.isbn9783642049217
dc.identifier.issn0302-9743
dc.identifier.urihttp://hdl.handle.net/10400.8/13037
dc.language.isoeng
dc.peerreviewedyes
dc.publisherSpringer
dc.relation.hasversionhttps://link.springer.com/chapter/10.1007/978-3-642-04921-7_38?utm_source=getftr&utm_medium=getftr&utm_campaign=getftr_pilot&getft_integrator=scopus
dc.relation.ispartofLecture Notes in Computer Science
dc.relation.ispartofAdaptive and Natural Computing Algorithms
dc.rights.uriN/A
dc.subjectSupport Vector Machine
dc.subjectIndependent Component Analysis
dc.subjectGeodesic Distance
dc.subjectGeneralize Regression Neural Network
dc.subjectNonlinear Dimensionality Reduction
dc.titleImproving Visualization, Scalability and Performance of Multiclass Problems with SVM Manifold Learningeng
dc.typebook part
dspace.entity.typePublication
oaire.citation.endPage379
oaire.citation.startPage370
oaire.citation.titleLecture Notes in Computer Science
oaire.citation.volume5495
oaire.versionhttp://purl.org/coar/version/c_970fb48d4fbd8a85
person.familyNameSilva
person.givenNameCatarina
person.identifier.orcid0000-0002-5656-0061
relation.isAuthorOfPublicationee28e079-5ca7-4842-9094-372c40f75c38
relation.isAuthorOfPublication.latestForDiscoveryee28e079-5ca7-4842-9094-372c40f75c38

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Improving visualization, scalability and performance of multiclass problems with SVM manifold learning.pdf
Size:
329.42 KB
Format:
Adobe Portable Document Format
Description:
We propose a learning framework to address multiclass challenges, namely visualization, scalability and performance. We focus on supervised problems by presenting an approach that uses prior information about training labels, manifold learning and support vector machines (SVMs). We employ manifold learning as a feature reduction step, nonlinearly embedding data in a low dimensional space using Isomap (Isometric Mapping), enhancing geometric characteristics and preserving the geodesic distance within the manifold. Structured SVMs are used in a multiclass setting with benefits for final multiclass classification in this reduced space. Results on a text classification toy example and on ISOLET, an isolated letter speech recognition problem, demonstrate the remarkable visualization capabilities of the method for multiclass problems in the severely reduced space, whilst improving SVMs baseline performance.
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.32 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

ESTG - Capítulos de livros