Table 3. RFMirTarget classification results on different feature subsets.
| Feature set | ACC (std) | SPE (std) | SEN (std) | MCC (std) |
| Cat 1: Alignment (2) | 59.34 (0.549) | 39.64 (1.066) | 73.31 (0.946) | 0.136 (0.011) |
| Cat 2: Thermodynamic (1) | 59.39 (1.156) | 45.38 (2.211) | 69.33 (1.051) | 0.150 (0.025) |
| Cat 3:Structural (5) | 67.57 (0.632) | 45.38 (0.562) | 83.31 (1.074) | 0.313 (0.013) |
| Cat 4: Seed (6) | 84.78 (0.407) | 82.98 (0.811) | 86.05 (0.537) | 0.687 (0.008) |
| Cat 5: Position-based (20) | 87.62 (0.462) | 84.67 (1.124) | 89.70 (0.314) | 0.744 (0.009) |
| Total (34) | 87.20 (0.434) | 85.84 (0.790) | 88.17 (0.604) | 0.737 (0.008) |
| Top ranked (12) | 89.53 (0.480) | 89.64 (0.673) | 89.46 (0.753) | 0.786 (0.009) |
The number of features in each category or set is given in parenthesis. Accuracy (ACC), specificity (SPE) and sensitivity (SEN) are expressed as percentages.