Research Article
Application of Deep Learning on the Prognosis of Cutaneous Melanoma Based on Full Scan Pathology Images
Table 2
The result of 5-fold cross-validation.
| Fold | Classification | Precision | Recall | F1-score | Support |
| Fold 1 valid | Living | 0.838 | 0.886 | 0.861 | 35 | Deceased | 0.000 | 0.000 | 0.000 | 6 | Fold 2 valid | Living | 0.892 | 0.943 | 0.917 | 35 | Deceased | 0.500 | 0.333 | 0.400 | 6 | Fold 3 valid | Living | 0.889 | 0.914 | 0.901 | 35 | Deceased | 0.400 | 0.333 | 0.364 | 6 | Fold 4 valid | Living | 0.838 | 0.912 | 0.873 | 34 | Deceased | 0.250 | 0.143 | 0.182 | 7 | Fold 5 valid | Living | 0.846 | 0.971 | 0.904 | 34 | Deceased | 0.500 | 0.143 | 0.222 | 7 |
|
|