autoMrP 1.0.6
- implements Deep MrP by Gopelrud as presented in https://doi.org/10.1017/S0003055423000035
- Set argument deep.mrp = TRUE to include Deep MrP in the ensemble
autoMrP 1.0.5
- drops missing values on y, L1.x, L2.x, L2.unit, L2.reg. Missing values on the DV would previously lead to errors in SVM
- works with continuous DV.
autoMrP 0.93
- block sampling in bootstrapping instead of state-stratified sampling
autoMrP 0.91
- bootstrapping returns GB prediction
- predictions do not fail if census data contains more factor levels than training data for SVM and Lasso
- svm post-stratification uses the user-specified formula instead of all information
- lasso post-stratification uses correct user-specified context level variables if L2.x and lasso.L2.x differ
- parallel processing loops are replicable now