Fig. 3From: A flexible approach for variable selection in large-scale healthcare database studies with missing covariate and outcome dataValidation plot of predicted probabilities of 3-year incidence of metabolic syndrome among middle-aged women using the SWAN data. The risk prediction models were the BART and XGBoost models with predictors selected via RR-BART, BI-BART and BI-XGB. missForest was used for imputationBack to article page