Research Article
Risk Assessment of Hip Fracture Based on Machine Learning
Table 16
Results of SVM with a Gaussian kernel.
| Gaussian | Mean | St. deviation | Best |
| Sensitivity (8 features) | 82.56% | 9.24% | 88.89% | Specificity (8 features) | 66.31% | 14.90% | 100.00% | Accuracy (8 features) | 75.74% | 8.23% | 93.54% | Sensitivity (12 features) | 96.22% | 5.05% | 100.00% | Specificity (12 features) | 58.08% | 15.09% | 76.92% | Accuracy (12 features) | 80.23% | 6.96% | 90.32% |
|
|