Figure 2.

Illustration of the correlation between actual times and predicted surgical times for spine surgery calculated by each model type: predicted times based on procedural averages and surgeon preference or customization, multivariable linear regression, random forest, bagging, and Extreme Gradient Boosting (XGBoost). The data set was split 80:20 (training:test), and the model was trained on the training set and validated on the test set.