Table 5 Comparison of SVM and neural network on testing data
From: Feature selection for gene prediction in metagenomic fragments
| Â | SVM | Neural network | ||||
|---|---|---|---|---|---|---|
Genomes | Sp | Sn | H.M. | Sp | Sn | H.M |
A. fulgidus | 96.46 | 87.26 | 91.61 | 95.60 | 82.09 | 88.33 |
M. jannaschii | 97.29 | 94.58 | 95.91 | 97.21 | 93.30 | 95.21 |
N. pharaonis | 97.37 | 82.71 | 89.44 | 96.10 | 77.27 | 85.66 |
B. aphidicola | 97.94 | 93.28 | 95.56 | 98.11 | 92.16 | 95.04 |
C. jeikeium | 97.31 | 88.64 | 92.77 | 97.15 | 84.84 | 90.58 |
C. tepidum | 95.93 | 80.84 | 87.74 | 94.71 | 75.99 | 84.32 |
H. pylori | 97.67 | 92.09 | 94.80 | 97.45 | 91.28 | 94.26 |
P. marinus | 98.58 | 87.65 | 92.79 | 98.71 | 85.22 | 91.47 |
W. endosymbiont | 88.10 | 89.66 | 88.87 | 88.69 | 87.04 | 87.86 |
B. pseudomallei | 97.56 | 85.83 | 91.32 | 97.95 | 81.43 | 88.93 |
P. aeruginosa | 97.64 | 88.88 | 93.05 | 97.70 | 86.90 | 91.98 |
Average | 96.53 | 88.31 | 92.17 | 96.31 | 85.23 | 90.33 |