Table 4 Performance of the models on training data with cross-validation
Indication | Model | Mean absolute percentage error (%) | Root mean squared error (min) | Mean absolute error (min) |
|---|---|---|---|---|
Hip (n = 810) | XGBoost | 23.73 (22.28–25.20) | 15.54 (14.26–16.81) | 10.89 (10.25–11.55) |
Linear multivariable regression | 40.40 (29.52–55.74) | 17.32 (16.01–18.63) | 12.33 (11.63–13.04) | |
Knee (n = 550) | XGBoost | 19.07 (17.57–20.60) | 14.92 (13.11–16.87) | 10.14 (9.38–10.91) |
Linear multivariable regression | 35.35 (27.64–45.30) | 19.26 (17.42–21.14) | 13.25 (12.30-14.25) |