Supervised pattern classification based on optimum-path forest

We present an approach for supervised classification, which interprets a training set as a complete graph, identifies prototypes in all classes, and computes an optimum-path forest rooted at them. The class of a sample in a tree is assumed to be the same of its root. A test sample is classified by identifying which tree would contain it. We show how to improve performance from the errors on an evaluation set, without increasing the training set. The advantages over others approaches are demonstrated using several experiments.