Clustering Multivariate Empiric Characteristic Functions for Multi-Class SVM Classification

  1. María-Dolores Cubiles-de-la-Vega 1
  2. Rafael Pino-Mejías 1
  3. Esther-Lydia Silva-Ramírez 2
  1. 1 Universidad de Sevilla
    info

    Universidad de Sevilla

    Sevilla, España

    ROR https://ror.org/03yxnpp24

  2. 2 Department of Language and Computer Sciences, University of Cadiz, Spain
Revista:
International Journal of Electrical and Computer Engineering

ISSN: 2313-3759

Año de publicación: 2012

Volumen: 6

Número: 4

Páginas: 380-384

Tipo: Artículo

Otras publicaciones en: International Journal of Electrical and Computer Engineering

Resumen

A dissimilarity measure between the empiric characteristic functions of the subsamples associated to the different classes in a multivariate data set is proposed. This measure can be efficiently computed, and it depends on all the cases of each class. It may be used to find groups of similar classes, which could be joined for further analysis, or it could be employed to perform an agglomerative hierarchical cluster analysis of the set of classes. The final tree can serve to build a family of binary classification models, offering an alternative approach to the multi-class SVM problem. We have tested this dendrogram based SVM approach with the one-against-one SVM approach over four publicly available data sets, three of them being microarray data. Both performances have been found equivalent, but the first solution requires a smaller number of binary SVM models.