An optimal multiclass classifier design
Marcelo Fiori, Matías Di Martino, Alicia Fernández
ICPR 2016 : 23rd International Conference on Pattern Recognition, Cancún Center, Cancún, México, 4-8 dec - 2016
Research Group(s): Tratamiento de Imagenes (gti)
Department(s): Procesamiento de Señales
Download the publication :

Abstract

The use of different evaluation measures for classification tasks have gained a significant amount of attention in the past decade, specially for those problems with multiple and imbalanced classes. However, the optimization of classifiers with respect to these measures is still heuristic, using ad-hoc rules with classical accuracy-optimized classifiers. We propose a classifier designed specifically to optimize one of the possible measures, namely, the so-called G-mean. Nevertheless, the technique is general, and it can be used to optimize generic evaluation measures. The optimization algorithm to train the classifier is described, and the numerical scheme is tested showing its usability and robustness. The code is publicly available, as well as the datasets used along this paper.

Additional data

""

BibTex references

Descargar BibTex bibtex

Other publications in the database

» Marcelo Fiori
» Matías Di Martino
» Alicia Fernández