Editing prototypes in the finite sample size case using alternative neighborhoods

Abstract

The recently introduced concept, of Nearest. Centroid Neighborhood is applied to discard outliers and prototypes in class overlapping regions in order to improve the performance of the Nearest Neighbor rule through an editing procedure. This approach is related to graph based editing algorithms which also define alternative neighborhoods in terms of geometric relations. Classical editing algorithms are compared to these alternative editing schemes using several synthetic and real data problems. The empirical results show that the proposed editing algorithm constitutes a good trade-off among performance and computational burden.