tune icon indicating copy to clipboard operation
tune copied to clipboard

add helper for bridging causal fits

Open simonpcouch opened this issue 2 years ago • 0 comments

PR 3/3, closes #652. Follows up on tidymodels/parsnip#955 and tidymodels/workflows#199.

library(tidymodels)
library(propensity)

propensity_wf <- workflow(Class ~ B, logistic_reg())
outcome_wf <- workflow(A ~ Class, linear_reg()) %>% add_case_weights(.wts)

res_tm <-
  fit_resamples(
    propensity_wf,
    resamples = bootstraps(two_class_dat, 1000),
    control = control_resamples(extract = identity)
  ) %>%
  weight_propensity(wt_ate, .treated = "Class2") %>%
  fit_resamples(
    outcome_wf,
    resamples = .,
    control = control_resamples(extract = tidy)
  )

res_tm %>%
  collect_extracts() %>%
  pull(.extracts) %>%
  bind_rows() %>%
  filter(term == "ClassClass2") %>%
  ggplot() +
  aes(x = estimate) +
  geom_histogram()
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Created on 2023-04-28 with reprex v2.0.2

simonpcouch avatar Apr 28 '23 20:04 simonpcouch