Clasificación mediante máquinas de soporte vectorial

Abstract

This article presents the main ideas and concepts behind Support Vector Machines. One of these concepts is the classifier or decisión hyperplane. Another key concept is the kernel considered as a way to measure similarity between objects of different class. However, Mercer's theorem is the mathematical basis for the functioning of Support Vector Machines, it allows to build classifiers in spaces whose dimension is greater than the original space. This allows to separate groups, which facilitates classification. Moreover, an application to simulated data is shown.