Algorithm | F-Measure |
|---|
| Â | T99-MRF | BL* | CG* | Turner99 |
|---|
ubcMEA
| 0.582 (0.574,0.591) | 0.680 (0.673,0.688) | 0.636 (0.628,0.644) | n/a |
ubcMFE
| 0.601 (0.592,0.609) | 0.686 (0.678,0.694) | 0.671 (0.663,0.679) | n/a |
rsMEA
| n/a | n/a | n/a | 0.623 (0.615,0.632) |
rsMFE
| n/a | n/a | n/a | 0.607 (0.598,0.615) |
gC-pMFmeas
| - | 0.711 (0.704, 0.718) | - | - |
- The table presents the prediction accuracy of different algorithms with different thermodynamic sets in terms of F-measure. The 95% percentile confidence intervals of their accuracies are also shown in parentheses. The parameter set T99-MRF refers to the Turner99 parameters in MultiRNAFold format. BL* and CG* are the parameter sets obtained by the BL and CG approaches of Andronescu et al. [9], respectively. Also, the Turner99 parameter set is the parameter set obtained by Mathews et al. [3]. "n/a" indicates cases in which a given algorithm is not applicable to a parameter set, since that does not match the energy model underlying the algorithm. The highest accuracies for MEA and MFE are shown in bold.
