[kkholst]

No description provided.

[kkholst]

@AndreasNordland, we need to discuss how to handle the id arguments in the most practical way...
Examples of the new mmrm methods can be found in inst/tinytest/test_mmrm.R

[kkholst]

@AndreasNordland, the following should now work:

set.seed(1)
dat <- targeted:::sim_cate(1e3) |>
  transform(z = rbinom(length(a), 1, 0.25),
            id = paste0("a", seq_along(a)))
rownames(dat) <- paste0("d", seq_len(nrow(dat)))
dat1 <- subset(dat, z==1)
g <- glm(u0 ~ w1 + a,
         data=dat1, family=binomial)
##
e1 <- estimate(g)
e <- c(
  'a=0'=expectation(predict, object=e1, data=transform(dat, a=0)),
  'a=1'=expectation(predict, object=e1, data=transform(dat, a=1))
)

id arguments can of course also be given directly to both the estimate and marginal functions (the above example automatically extracts ids from the row names of the data / model.frame.

[github-actions]

Github pages

Review the pkgdown webpage for the PR here