A supervised classifier scheme based on clustering algorithms

A. Hernandez-Matamoros, E. Escamilla-Hernandez, K. Perez-Daniel, M. Nakano-Miyatake, H. Perez-Meana

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

This paper proposes a new classifier scheme based on classical clustering algorithms, such as the Batchelor & Wilkins y K-means algorithms which are trained in a similar form that the artificial neural network (ANN) or support vector machines (SVM). Proposed scheme has the advantage that if a new class is added, it is not necessary to train he classifier completely, but only add a new class. Experimental results show that the proposed scheme provides classification rates quite similar to those provided by the SVM with much less computational complexity.

Original languageEnglish
Title of host publicationProceedings of the 2014 IEEE Central America and Panama Convention, CONCAPAN 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781479975846
DOIs
StatePublished - 30 Dec 2014
Event2014 34th IEEE Central America and Panama Convention, CONCAPAN 2014 - Panama City, Panama
Duration: 12 Nov 201414 Nov 2014

Publication series

NameProceedings of the 2014 IEEE Central America and Panama Convention, CONCAPAN 2014

Conference

Conference2014 34th IEEE Central America and Panama Convention, CONCAPAN 2014
Country/TerritoryPanama
CityPanama City
Period12/11/1414/11/14

Keywords

  • Pattern recognition
  • Supervised training
  • self-organizing maps
  • support vector machines

Fingerprint

Dive into the research topics of 'A supervised classifier scheme based on clustering algorithms'. Together they form a unique fingerprint.

Cite this