A new classification rule based on nearest neighbour search
Francisco Moreno-seco, Jose Oncinadept, Lenguajes Sistemas Inform'aticos · 2008
Abstract The nearest neighbour (NN) classification rule is usuallychosen in a large number of pattern recognition systems due to its simplicity and good behaviour. As the problemof finding the nearest neighbour of an unknown sample is also of interest in other scientific communities (very largedatabases, data mining, computational geometry,...), avast number of fast nearest neighbour search algorithms have been developed during the last years. In order to im-prove classification rates, the k-NN rule is often used in-stead of the NN rule, but it yields higher classification times. In this work we introduce a new classification rule applica-ble to many of those algorithms in order to obtain classification rates better than those of the nearest neighbour (similarto those of the