This repository contains R programs for the article, “More powerful goodness-of-fit tests for uniform stochastic ordering,” by Dewei Wang, Chuan-Fa Tang, and Joshua Tebbs. This article has been submitted for publication.
Prior to using R programs on this repository, please install R package "TestUSO" (https://github.com/Harrindy/TestUSO) using the following:
library(devtools)
install_github("Harrindy/TestUSO",force=TRUE)
To repeat our simulation:
1. Open Simulation.R in R (Rstudio).
3. Follow the comments to change parameters to the ones in the simulation setting that you want to repeat.
4. Make sure that "TestUSO" package has been installed and R (or Rstudio) can access the internet.
5. Run Simulation.R.
The premature infant data is provided in PrematureInfant.CSV. Due to a confidentiality issue, we cannot provide the original data. However, because the ODC and all the methods considered in the paper are invariant to an increasing transformation on the two samples, we divided each observation by the maximum of the two samples and presented the transformed data in PrematureInfant.csv.
To reproduce the analysis of the premature infant data:
1. Make sure that "TestUSO" package has been installed and R (or Rstudio) can access the internet.
2. Run PrematureInfant.R.