Title: qape: R package to estimate prediction accuracy in mixed models based on bootstrap methods
Authors: Alicja Wolny-Dominiak - University of Economics in Katowice (Poland) [presenting]
Tomasz Zadlo - University of Economics in Katowice (Poland)
Abstract: In order to compare relevance of two models, the coefficient of determination, residual variance or AIC are typically calculated. One of the most popular prediction accuracy measures is the mean squared error of prediction (MSEP) and its different modifications. The alternative is the quantile absolute prediction error (QAPE) reflects the relation between the magnitude of the error and the probability of its realization. The R package qape to estimate such errors is presented. The parametric and residual bootstrap procedures, as well as the double bootstrap, are implemented. The LMM, GLMM and GLM models are applied.