Table 4.
Nearest neighbor performance (10-fold cross-validation error, %)
Dataset | Spectrum | Hamming | Kimura | Smith-Waterman |
ACG | 2.49 ± 0.87 | 11.44 ± 1.52 | 5.51 ± 0.86 | 3.66 ± 0.66 |
Hesperiidae | 3.57 ± 1.08 | 14.49 ± 2.36 | 3.81 ± 1.26 | 5.45 ± 1.20 |
Astraptes | 1.07 ± 1.81 | 3.61 ± 2.77 | 1.71 ± 1.96 | 1.64 ± 1.03 |
Bats Guyana | 1.63 ± 1.22 | 2.72 ± 1.83 | 1.63 ± 1.22 | 1.63 ± 1.22 |
Birds of North America | 6.22 ± 1.50 | 18.38 ± 2.05 | 6.02 ± 1.36 | 8.20 ± 1.53 |
Fish Australia | 5.50 ± 3.27 | 5.87 ± 4.01 | 5.35 ± 3.36 | 5.35 ± 3.36 |
Fish larvae† | 2.86 | 11.43 | 8.57 | 5.71 |
† leave-one-out validation error is reported