An Immune Algorithm that Uses a Master Cell for Component Selection for the kNN Method

Abstract

This paper details an immune algorithm (IA) that improves the accuracy of a kNN (k-Nearest Neighbor) method when using it for prognosis and diagnosis. Our IA uses a master cell to produce new groups of immune candidate cells. The IA selects the best components that must be used by the kNN method. The best combinations of components found by our IA helped to improve the average accuracy of the kNN method by 1% for the UCI breast cancer data (prognosis) and almost 3% for the UCI Cleveland heart (diagnosis) data set.