Weighted k-NN
1) for the segmentation data 3-WNN method gives 144 classification errors in comparsion to 187 errors given by classical 3-NN method. This improves classification by 2 % (from 91% to 93%)
2) for the ionosphere data 3-WNN method gives 5 classification errors in comparsion to 7 errors given by classical 3-NN method. This improves classification approximately by 1% (from 95 to 96 %)
3) for the monk3 data 3-WNN method gives 17 classification errors in comparsion to 18 errors given by classical 3-NN method. This slightly improves classification.
4) for the sonar data 3-WNN method gives 6 classification errors in comparsion to 16 errors given by classical 3-NN method. This improves classification by 10% (from 84% to 94%)
5) for the monk2 data 3-WNN method gives 137 classification errors in comparsion to 149 errors given by classical 3-NN method. This improves classification by 3 % (from 65% to 68%)