Table 5.
Performances of the six machine learning classifiers for predicting molecular subtype.
| Classifier | Subtype | Precision | Recall | F1-score | Overall Accuracy |
|---|---|---|---|---|---|
| MLP | TNBC | 0.813 | 0.520 | 0.634 | 0.735 |
| HR+/HER2− | 0.771 | 0.661 | 0.712 | ||
| HER2+ | 0.704 | 0.852 | 0.771 | ||
| GPC | TNBC | 0.692 | 0.360 | 0.474 | 0.623 |
| HR+/HER2− | 0.628 | 0.482 | 0.545 | ||
| HER2+ | 0.613 | 0.802 | 0.695 | ||
| LDA | TNBC | 0.714 | 0.400 | 0.513 | 0.642 |
| HR+/HER2− | 0.674 | 0.518 | 0.586 | ||
| HER2+ | 0.619 | 0.802 | 0.699 | ||
| SVM | TNBC | 0.778 | 0.280 | 0.412 | 0.623 |
| HR+/HER2− | 0.697 | 0.411 | 0.517 | ||
| HER2+ | 0.592 | 0.877 | 0.707 | ||
| RF | TNBC | 0.625 | 0.400 | 0.488 | 0.636 |
| HR+/HER2− | 0.628 | 0.482 | 0.545 | ||
| HER2+ | 0.641 | 0.815 | 0.718 | ||
| LR | TNBC | 0.733 | 0.440 | 0.550 | 0.679 |
| HR+/HER2− | 0.700 | 0.625 | 0.660 | ||
| HER2+ | 0.660 | 0.790 | 0.719 |
SVM, Support Vector Machine (radial bias function); RF, Random Forest; LR, Logistic Regression; LDA, Linear Discriminant Analysis; GPC, Gaussian Process Classifier; MLP, Multilayer Perceptron; TNBC, triple-negative breast cancer; HR, hormone receptor; HER2, human epidermal growth factor receptor-2.