Table 2 Model Characteristics

Neural Network

0.05

CV: 0.840 [0.825–0.855]

DL: 0.841 [0.816–0.863]

72.9%

[69.1–76.7%]

77.5%

[76.2–78.7%]

3.25

[3.03–3.47]

15.1%

[14.2–16.0%]

98.1%

[97.9–98.4%]

XGBoost

0.25

CV: 0.852 [0.839–0.865]

DL: 0.851 [0.827–0.874]

80.6%

[77.1–84.1%]

73.7%

[72.4–74.9%]

3.08

[2.87–3.29]

14.4%

[13.5–15.3%]

98.6%

[98.3–98.8%]

Clinician-Guided Regression

0.05

CV: 0.746 [0.718–0.775]

DL: 0.763 [0.734–0.793]

69.1%

[62.9–75.4%]

65.5%

[64.3–66.7%]

2.01

[1.79–2.23]

9.0%

[7.2–10.9%]

97.4%

[96.9–98.0%

ML Hybrid Regression

0.32

CV: 0.810 [0.787–0.832]

DL: 0.824 [0.800–0.849]

74.7%

[69.8–79.6%]

73.5%

[72.1–74.9%]

2.84

[2.46–3.09]

13.9%

[12.7–15.1%]

98.1%

[97.7–98.4%]

AWOL-S^a

0.05

DL: 0.762 [0.713–0.812]

78.2%

[66.0–89.3%]

60.0%

[57.0–63.0%]

1.95

[1.62–2.28]

9.4%

[6.8–12.3%]

98.1%

[96.8–99.1%]