Table 4. Performance summary of the best performing feature-based classifiers (all with RF feature selection) as well as of the four scalar metrics from univariate logistic regression.
Accuracy, sensitivity, specificity and AUC were reported based on the 58 separate injury predictions in the leave-one-out cross-validation framework. The average AUC measures (and 95% CI) for the training datasets were also reported.
| Deep learning | SVM | RF (95% CI) | BrIC | CSDM-WB | CSDM-CC | Peak-CC | |
|---|---|---|---|---|---|---|---|
| Accuracy | 0.862 | 0.828 | 0.842 (0.810–0.862) | 0.776 | 0.741 | 0.776 | 0.690 |
| Sensitivity | 0.840 | 0.760 | 0.787 (0.760–0.840) | 0.640 | 0.640 | 0.760 | 0.600 |
| Specificity | 0.879 | 0.879 | 0.883 (0.849–0.909) | 0.879 | 0.818 | 0.788 | 0.758 |
| AUC-Testing | 0.892 | 0.872 | 0.856 | 0.781 | 0.786 | 0.771 | 0.737 |
|
AUC-Training average (95% CI) |
0.967 (0.933, 0.978) |
0.963 (0.951, 0.981) |
1.000 (1.000, 1.000) |
0.805 (0.797, 0.831) | 0.838 (0.831, 0.860) |
0.815 (0.807, 0.843) |
0.770 (0.760, 0.791) |