Figure 3.

Prediction accuracy measured in $\mathit{MSE}$ using the XGBoost method for predicting scaled sound pressure in the Airfoil dataset under various missing rates. The $MSE$ is estimated based on a five-fold cross-validation procedure on the imputed dataset. The dotted lines around the solid curves refer to $95\%$ empirical Monte-Carlo intervals. The solid lines are Monte-Carlo means of the cross-validated $MSE$. The horizontal dotted line in red refers to the cross-validated $MSE$ of the XGBoost method fitted to the Airfoil dataset without any missing values.