Improving nearest neighbor classification using Ensembles of Evolutionary Generated Prototype Subsets

Prototype selection reduces the dataset before the application of a classifier in order to achieve an improved accuracy and/or a considerable reduction in the number of instances. Among the proposed algorithms, evolutionary methods are the state-of-theart. In (Verbiest et al., 2016), we developed a framework to further enhance the performance of these methods with minimal additional effort. We recall our findings here.