Table 6.
The prediction performance of the automated classifier, yasMir, miPred, tripletSVM and our method evaluated the same testing data set
| Test sets | Automated classifier (11) | yasMiR (10) | miPred (7) | Triplet SVM (6) | Our method |
|---|---|---|---|---|---|
| TE-H | 94.30 | 93.77 | 93.50 | 87.96 | 98.37 |
| IE-NH | 94.91 | 94.11 | 95.64 | 86.15 | 95.31 |
| IE-NC | 77.71 | 82.95 | 68.68 | 78.37 | 83.22 |
| IE-M | 96.77 | 100 | 87.09 | 0 | 100 |
TE-H (123 human pre-miRNA and 246 pseudo hairpins), IE-NH (1918 pre-miRNA across 40 non-human species and 3836 pseudo hairpins), IE-NC (12 387 functional ncRNAs) and IE-M (31 mRNAs). The values are percentages of correct prediction for each method on each data set. The highest values are in bold.