An R package for the paper “Wasserstein F-tests and confidence bands for the Frechet regression of density response curves”.
You can install the released version of WRI from CRAN with:
This is a basic example which shows you how to solve a common problem:
library(WRI)
data(strokeCTdensity)
predictor = strokeCTdensity$predictors
dSup = strokeCTdensity$densitySupport
densityCurves = strokeCTdensity$densityCurve
xpred = predictor[3, ]
res = wass_regress(rightside_formula = ~., Xfit_df = predictor,
Ytype = 'density', Ymat = densityCurves, Sup = dSup)
# compute the density band for the third observation
confidence_Band1 = confidenceBands(res, Xpred_df = xpred, type = 'density')
strokeCTdensity
: clinical, radiological scalar variables and density curves of the hematoma of 393 stroke patientswass_regress
: perform Frechet Regression with the Wasserstein Distancewass_R2
: compute Wasserstein coefficient of determinationglobalFtest
: perform global F test for Wasserstein regressionpartialFtest
: perform partial F test for Wasserstein regressionsummary.WRI
: provide summary information of Wasserstein regressionconfidenceBands
: compute intrinsic confidence bands and density bands