Fast K-nearest-neighbour search algorithm for nonparametricclassification
SeongJoon Baek, Koeng-Mo Sung · Electronics Letters · 2000
A fast KNN search algorithm for nonparametric classification is presented. The proposed algorithm uses a projection vector to accelerate the classification process by eliminating the need to calculate a large number of distances. The algorithm also uses a linked list to efficiently retain the intermediate k closest vector and is combined with the PDS technique to obtain further acceleration. Simulation results confirm the effectiveness of the proposed algorithms.