Table 4.
Sensitivity, Specificity, and Balanced Accuracies for Patients with Mild Head Injury with at Least One Computer Tomography Abnormality versus Patients with Mild Head Injury with No Computer Tomography Abnormalities. Partial Least Squares-Discriminant Analysis, Random Forest, and Support Vector Machine Classification Models with 95% Confidence Intervals Included
| Model | Sensitivity (%) |
Specificity (%) |
Balanced accuracy (%) |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Mean | SD | 95% CI | Mean | SD | 95% CI | Mean | SD | 95% CI | |
| PLS-DA | 80.2 | 9.2 | ±2.5 77.7–82.7 |
33.2 | 14.4 | ±4.0 29.2–37.2 |
56.7 | 7.2 | ±2.0 54.7–58.7 |
| RF | 62.5 | 14.1 | ±3.9 58.6–66.4 |
46.6 | 15.3 | ±4.2 42.4–50.8 |
54.5 | 8.5 | ±2.3 52.2–56.8 |
| SVM | 61.1 | 13.4 | ±3.7 57.4–64.8 |
42.8 | 17.9 | ±4.9 37.9–47.7 |
52.0 | 10.0 | ±2.7 49.3–54.7 |
SD, standard deviation; CI, confidence interval; PLS-DA, partial least squares-discriminant analysis; RF, random forest; SVM, support vector machine.