Additional file 1 of Prediction models developed using artificial intelligence: similar predictive performances with highly varying predictions for individuals – an illustration in deep vein thrombosis

Supplementary Material 1: Table S1. Details on model development [22, 36–38]. Table S2. Tuning of parameters based on repeated cross-validation. Table S3. Risk probabilities and prediction interval for ten random individuals. Table S4. Calibration statistics (95% Confidence Intervals) for five models. Figure S1: ICE Plot of age for the 5 different models. Figure S2. ICE Plot of sex for the 5 different models. Figure S3. ICE Plot of history of previous DVT for the 5 different models. Figure S4. ICE Plot of dichotomized d-dimer value for the 5 different models. Figure S5. ICE Plot of active malignancy for the 5 different models. Figure S6. ICE Plot of alternative diagnosis for the 5 different models.