From: Multivariate classification of urine metabolome profiles for breast cancer diagnosis
 | Actual | Output | Accuracy | Sensitivity | Specificity | |
---|---|---|---|---|---|---|
 |  | Cancer | Normal |  |  |  |
Decision Tree | Â | Â | Â | Â | Â | Â |
Confidence=0.25 | Cancer | 46 | 4 | 90.06% | 92.00% | 88.00% |
Pruning=true | Normal | 6 | 44 | Â | Â | Â |
Random Forest | Â | Â | Â | Â | Â | Â |
Tree=500 | Cancer | 46 | 4 | 91.02% | 92.00% | 90.00% |
Feature=6 | Normal | 5 | 45 | Â | Â | Â |
Support Vector Machine (Linear) | Â | Â | Â | Â | Â | Â |
Cost=25 | Cancer | 48 | 2 | 91.41% | 96.00% | 86.00% |
Gamma=0.33 | Normal | 7 | 43 | # of Support Vectors: 26 | ||
Support Vector Machine (Gaussian) | Â | Â | Â | Â | Â | Â |
Cost=45 | Cancer | 46 | 4 | 91.02% | 92.00% | 90.00% |
Gamma=0.33 | Normal | 5 | 45 | # of Support Vectors: 26 | ||
Classification results for the second feature set (Table 1B) |