Skip to main content

Table 3 Performance of the models on test data

From: Leveraging machine learning for duration of surgery prediction in knee and hip arthroplasty – a development and validation study

Indication

Model

Mean absolute percentage error (%)

Root mean squared error (min)

Mean absolute error (min)

Hip (n = 526)

XGBoost

22.66 (21.10-24.28)

17.18 (15.42-19.00)

12.13 (11.27–13.02)

Linear multivariable regression

23.89 (22.16–25.72)

17.68 (15.94–19.50)

12.51 (11.63–13.44)

Knee (n = 400)

XGBoost

23.30 (19.96–23.61)

19.03 (16.93–21.39)

13.61 (12.56–14.73)

Linear multivariable regression

21.51 (19.66–23.49)

19.24 (17.04–21.64)

13.55 (12.45–14.69)

  1. 95% confidence intervals in parenthesis. Confidence intervals were derived using bootstrapping with 10,000 repetitions
  2. A) Hip arthroplasty