Table 9.
Specificity of our ensemble when applied to the negative testing data, compared with yasMir (the 2nd best sensitivity from Table 8)
| Negative data | No. of sequences | Our method |
yasMir |
||
|---|---|---|---|---|---|
| Correctly classified (%) | FPR (%) | Correctly classified (%) | FPR (%) | ||
| Pseudo hairpin | 4494 | 93.74 | 6.26 | 86.91 | 13.09 |
| Shuffle | 21 470 | 88.35 | 11.65 | 83.69 | 16.31 |
| IE-NC (1238ncRNA) | 12 387 | 83.22 | 16.78 | 82.95 | 17.05 |
| Average | 12 784 | 88.44 | 11.56 | 84.52 | 15.48 |
Correctly classified (%) is the percent of the correctly classified as not pre-miRNAs, FPR (%) is the false positive rate.