Table 2 Prediction performance metrics for two classifiers: RFRI and single-layer FNN
From: Algorithms of the Möbius function by random forests and neural networks
Metric | Formula | Explanation | RFRI | FNN |
|---|---|---|---|---|
Accuracy | \(\frac{\text {TP+TN}}{\text {TP+TN+FP+FN}}\) | Percentage of correct classifications | 0.9493 | 0.7871 |
TPR/Sensitivity/Recall | \(\frac{\text {TP}}{\text {TP+FN}}\) | Rate of correctly classified positives | 0.5865 | 0.9477 |
FPR | \(\frac{\text {FP}}{\text {FP+TN}}\) | Rate of incorrectly classified positives | 0.0229 | 0.2248 |
Precision | \(\frac{\text {TP}}{\text {TP+FP}}\) | Fraction of positive predictions thatwere actually positives | 0.6626 | 0.2384 |
\(F_1\)-Score | \(\frac{2\cdot \text {Precision} \cdot \text {Recall}}{\text {Precision} + \text {Recall}}\) | Harmonic mean of the precision and recall | 0.6223 | 0.3809 |