This code is an accompaniment to the paper titled "Which Wilcoxon should we use? An interactive rank test and other alternatives", and produce all plots in the paper.
The i-Wilcoxon test is for two-sample comparison with unpaired data in a randomized trial that allows human interaction. We also investigate several non-interactive variants of the Wilcoxon signed-rank tests.
To reproduce the experiments:
$Rscript reproduce.R
To generate the figures:
$Rscript plots.R
tests.R: the i-Wilcoxon test; variants of the Wilcoxon test; linear-CATE-test.
sim_dat.R: simulate outcomes of different types.
result/.: p-values of each method in each experiment.
The code was tested using R (version 3.6.0) and the following packages: magrittr, reshape2, ggplot2, randomForest, ash, caTools, tibble, dplyr, tidyr, doParallel, foreach, energy, MASS