Table 3 The features of the classifiers based on machine leaning models
Features | Accuracy (95% CI) | Specifity (95% CI) | Sensitivity (95% CI) |
|---|---|---|---|
5 Mb windows-RF | 0.712 (0.703–0.720) | 0.753 (0.736–0.771) | 0.642 (0.625–0.659) |
Sub-compartments-RF | 0.887 (0.880–0.893) | 0.901 (0.891–0.911) | 0.858 (0.846–0.870) |
Promoter profiles-RF | 0.953 (0.948–0.957) | 0.943 (0.937–0.950) | 0.973 (0.968–0.978) |
5 Mb windows-LR | 0.716 (0.707–0.724) | 0.728 (0.711–0.746) | 0.692 (0.675–0.709) |
Sub-compartments-LR | 0.833 (0.826–0.840) | 0.822 (0.810–0.833) | 0.856 (0.843–0.869) |
Promoter profiles-LR | 0.880 (0.874–0.887) | 0.870 (0.861–0.880) | 0.901 (0.891–0.911) |
5 Mb windows-SVM | 0.538 (0.527–0.550) | 0.582 (0.550–0.614) | 0.449 (0.418–0.480) |
Sub-compartments-SVM | 0.829 (0.823–0.836) | 0.936 (0.929–0.943) | 0.614 (0.598–0.630) |
Promoter profiles-SVM | 0.918 (0.913–0.923) | 0.956 (0.950–0.961) | 0.841 (0.829–0.854) |