Research Article
Risk Assessment of Hip Fracture Based on Machine Learning
Table 14
Results of SVM with a sigmoid kernel and posterior probability regions.
| Sigmoid + PPR | Mean | St. deviation | Best |
| Sensitivity (8 features) | 55.39% | 12.37% | 77.78% | Specificity (8 features) | 59.23% | 14.40% | 84.62% | Accuracy (8 features) | 57.00% | 10.09% | 80.65% | Sensitivity (12 features) | 55.94% | 11.23% | 83.33% | Specificity (12 features) | 58.38% | 15.66% | 69.23% | Accuracy (12 features) | 56.97% | 8.67% | 77.41% |
|
|