Skip to main content

Advertisement

Table 1 Results simulation study

From: Prediction intervals for future BMI values of individual children - a non-parametric approach by quantile boosting

95% PIs p= 10 p= 500
  mboost quantregForest mboost quantregForest
Linear setup     
π ^ | x 1 0.9454 0.9948 0.9361 0.9997
π ^ | x 2 0.9489 0.9689 0.9425 0.9889
π ^ | x 3 0.9466 0.9561 0.9418 0.9609
π ^ | x 4 0.9437 0.9307 0.9400 0.9471
π ^ | x 5 0.9405 0.9310 0.9373 0.9534
Non-linear setup     
π ^ | x 1 0.9486 0.9721 0.9662 0.9832
π ^ | x 2 0.9494 0.9925 0.9623 0.9961
π ^ | x 3 0.9490 0.9940 0.9521 0.9954
π ^ | x 4 0.9460 0.9785 0.9407 0.9792
π ^ | x 5 0.9314 0.8743 0.9171 0.8942
  1. Mean conditional coverage resulting from 95% PIs for both setups and both scenarios. In every row, the value of the better performing algorithm (with the mean conditional coverage closer to the expected coverage of 95%) for each setup is printed in bold.