The k-Nearest Neighbours classifier [43] (kNN) consists
of the assignment of an unclassified vector using the closest 
vectors found in the training set. In this classifier, the
Euclidean distance is used. Due to the fact that kNN is based on
distances between sample points in the feature space, features
need to be normalized to avoid that some features are weighted
more strongly than others. Hence, all features have been
normalized to unit variance and zero mean. Moreover, kNN presents
another inherent problem, which is the uniform weighting of
features regardless their discriminant power. In order to solve
this problem we have included a feature selection step which
automatically selects the set of the most discriminant features.
Here we have used the Sequential Forward Selection (SFS)
algorithm [87], which is a widely known technique that
selects a local optimum solution in a computationally attractive
way. SFS starts by selecting the best single feature and, in an
iterative process, subsequent features are selected one at a time
which in combination with the already selected ones, maximizes an
Euclidean distance based criterion function.