Table 1 Performances of triplet-SVM, miPred and miR-KDE in predicting human pre-miRNAs.
From: Using a kernel density estimation based classifier to predict species-specific microRNA precursors
| Â | %SE | %SP | %ACC | %Fm | %MCC |
|---|---|---|---|---|---|
Five-fold cross-validation on HU400 | |||||
   triplet-SVM | 86.5% | 91.5% | 89.0% | 88.7% | 78.1% |
   miPred | 87.5% | 98.0% | 92.8% | 92.3% | 86.0% |
   miR-KDE | 90.5% | 97.5% | 94.0% | 93.8% | 88.2% |
Using HU400 to predict HU216 | |||||
   triplet-SVM | 83.3% | 86.1% | 84.7% | 84.5% | 69.5% |
   miPred | 88.0% | 88.0% | 88.0% | 88.0% | 75.9% |
   miR-KDE | 88.9% | 92.6% | 90.7% | 90.6% | 81.5% |