Table 3 The jackknife results for the method of support vector machine trained by the six combinations listed in table 1. Three attributes have been defined, namely binary system (denoted by "A" in the table), thermodynamic profile ("B" in the table) and composition ("C" in the table). Seven combinations of the attributes are put forward, which are A+B+C (means "binary, thermodynamic and composition"), A+B (means "binary and thermodynamic"), B+C (means "thermodynamic and composition"), A+C (means "thermodynamic and composition"), A (means "binary only"), B (means "thermodynamic only") and C (means "composition"). The self-consistency and jackknife test are executed in all the seven vector space respectively to compare the contribution from each of the three attributes. To save space, here we just listed the results of jackknife test. This table lists results of Dieter's dataset. See table 4 for Satron's dataset. Self-consistency results have been placed in the supplemental file (see additional file 3) Dieter's dataset, jackknife test:
From: Demonstration of two novel methods for predicting functional siRNA efficiency
| Â | A+B+C | A+B | B+C | A+C | A | B | C |
|---|---|---|---|---|---|---|---|
Cut-off 0.7 | Â | Â | Â | Â | Â | Â | Â |
Accuracy | 94.78% | 94.90% | 85.97% | 94.86% | 96.13% | 78.69% | 81.65% |
Sensitivity | 86.51% | 87.87% | 67.98% | 87.19% | 91.14% | 48.64% | 57.90% |
Specificity | 98.35% | 97.94% | 93.75% | 98.17% | 98.29% | 91.69% | 91.93% |
Pearson | 0.9726 | 0.9752 | 0.8522 | 0.9749 | 0.9808 | 0.7189 | 0.7377 |
ROC | 0.9899 | 0.9922 | 0.9302 | 0.9913 | 0.9952 | 0.8411 | 0.8809 |
Cut-off 0.6 | Â | Â | Â | Â | Â | Â | Â |
Accuracy | 94.65% | 96.01% | 83.83% | 95.80% | 96.71% | 76.31% | 80.09% |
Sensitivity | 94.32% | 96.19% | 82.37% | 95.42% | 96.61% | 73.39% | 79.66% |
Specificity | 94.96% | 95.84% | 85.21% | 96.16% | 96.80% | 79.06% | 80.50% |
Pearson | 0.9735 | 0.9786 | 0.8469 | 0.9775 | 0.9825 | 0.7181 | 0.7619 |
ROC | 0.9912 | 0.9947 | 0.9223 | 0.9937 | 0.9967 | 0.8436 | 0.885 |
Cut-off 0.5 | Â | Â | Â | Â | Â | Â | Â |
Accuracy | 94.65% | 95.56% | 83.67% | 95.23% | 96.42% | 77.46% | 79.47% |
Sensitivity | 96.53% | 97.10% | 90.85% | 96.85% | 97.79% | 88.71% | 86.75% |
Specificity | 91.13% | 92.67% | 70.21% | 92.20% | 93.85% | 56.38% | 65.84% |
Pearson | 0.9726 | 0.974 | 0.8415 | 0.9761 | 0.98 | 0.7172 | 0.741 |
ROC | 0.9906 | 0.9928 | 0.9121 | 0.9926 | 0.9951 | 0.8435 | 0.8668 |