s1) Stub - Probablistic index models

Nathan Brouwer

2018-12-11

These are notes for a tutorial that hasn’t been written

Probalistic index models

https://cran.r-project.org/web/packages/pim/index.html

library(pim)

wb <- aggregate(warpbreaks$breaks,
                by = list(w = warpbreaks$wool,
                          t = warpbreaks$tension),
                FUN = mean)


comp <- expand.grid(1:nrow(warpbreaks), 
                    1:nrow(warpbreaks))
comp <- comp[warpbreaks$tension[comp[,1]] == warpbreaks$tension[comp[,2]],] # only compare within blocks
m <- pim(breaks ~ wool, data = warpbreaks, compare = comp, link = "identity",  vcov.estim = score.vcov)
summary(m)
friedman.test(x ~ w | t, data = wb)