Table 2 Comparison of participant profiles in non-OSA and OSA groups.
From: Enhancing construction safety: predicting worker sleep deprivation using machine learning algorithms
Models | (95% C.I) AUC | (95% C.I) Accuracy (%) | (95% C.I) Sensitivity (%) | (95% C.I) Specificity (%) | (95% C.I) PPV (%) | (95% C.I) PNV (%) |
|---|---|---|---|---|---|---|
Dataset training | ||||||
 LR | 0.92 (0.88–0.95) | 84.23 (79.16–88.28) | 81.93 (74.67–87.38) | 86.54 (79.19–91.72) | 85.84 (80.11–90.31) | 82.77 (77.25–87.23) |
 DT | 0.98 (0.99–1.00) | 99.04 (97.18–99.41) | 97.72 (93.51–99.31) | 98.35 (97.40–98.89) | 98.33 (95.42–98.90) | 97.78 (92.73–99.62) |
 SVM | 0.97 (0.93–0.99) | 91.78 (88.96–94.92) | 91.78 (86.23–95.78) | 91.80 (86.12–95.91) | 91.78 (85.65–95.21) | 91.78 (85.65–95.21) |
 RF | 0.98 (0.97–0.99) | 98.05 (96.45–99.27) | 97.72 (93.51–99.88) | 98.38 (94.51–98.43) | 98.36 (93.32–98.04) | 97.74 (92.59–98.61) |
Dataset test | ||||||
 LR | 0.85 (0.79–0.92) | 76.02 (65.47–84.91) | 71.50 (58.5–82.7) | 87.88 (66.6–98.4) | 89.50 (77.74–96.88) | 54.20 (35.47–71.29) |
 DT | 0.88 (0.78–0.95) | 84.45 (74.26–91.60) | 81.41 (69.3–88.7) | 87.88 (67.5–98.5) | 89.61 (81.15–94.25) | 70.60 (52.89–83.58) |
 SVM | 0.83 (0.75–0.89) | 79.61 (68.26–86.79) | 71.50 (59.6–82.8) | 87.88 (67.5–98.4) | 87.51 (79.82–92.78) | 61.02 (45.51–75.10) |
 RF | 0.81 (0.76–0.90) | 76.01 (68.41–84.88) | 79.68 (67.4–89.4) | 74.84 (52.3–88.5) | 86.81 (78.61–92.34) | 54.60 (40.65–68.50) |
