In the last ten years, different approaches as for example, SVMs have
been proposed. They are generally non parametric classifiers which
show a great ability to optimize classification issues, minimizing the
empirical classification errors while maximizing the class separations.