Performance evaluation of prototype selection algorithms for nearest neighbor classification
J. Salvador Sánchez, R. Barandela, R. Alejo, Ana I. Marqués · 2002
Prototype selection is primarily effective in improving the classification performance of nearest neighbor (NN) classifier and also partially in reducing its storage and computational requirements. This paper reviews some prototype selection algorithms for NN classification and experimentally evaluates their performance using a number of real data sets. Finally, new approaches based on combining the NN and the nearest centroid neighbor (NCN) of a sample are also introduced.