Skip to content

Updates to Vignette #179

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Feb 13, 2023
Merged

Updates to Vignette #179

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 9 additions & 8 deletions vignettes/getstarted.Rmd
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ dml_data_bonus = DoubleMLData$new(df_bonus,
print(dml_data_bonus)

# matrix interface to DoubleMLData
dml_data_sim = double_ml_data_from_matrix(X=X, y=y, d=d)
dml_data_sim = double_ml_data_from_matrix(X = X, y = y, d = d)
dml_data_sim
```

Expand All @@ -94,12 +94,12 @@ library(mlr3learners)
# surpress messages from mlr3 package during fitting
lgr::get_logger("mlr3")$set_threshold("warn")

learner = lrn("regr.ranger", num.trees=500, mtry=floor(sqrt(n_vars)), max.depth=5, min.node.size=2)
ml_g_bonus = learner$clone()
learner = lrn("regr.ranger", num.trees = 500, max.depth = 5, min.node.size = 2)
ml_l_bonus = learner$clone()
ml_m_bonus = learner$clone()

learner = lrn("regr.glmnet", lambda = sqrt(log(n_vars)/(n_obs)))
ml_g_sim = learner$clone()
ml_l_sim = learner$clone()
ml_m_sim = learner$clone()
```

Expand All @@ -111,9 +111,10 @@ When initializing the object for PLR models `DoubleMLPLR`, we can further set pa
* The number of folds used for cross-fitting `n_folds` (defaults to `n_folds = 5`) as well as
* the number of repetitions when applying repeated cross-fitting `n_rep` (defaults to `n_rep = 1`).

Additionally, one can choose between the algorithms `"dml1"` and `"dml2"` via `dml_procedure` (defaults to `"dml2"`). Depending on the causal model, one can further choose between different Neyman-orthogonal score / moment functions. For the PLR model the default score is `"partialling out"`.
Additionally, one can choose between the algorithms `"dml1"` and `"dml2"` via `dml_procedure` (defaults to `"dml2"`). Depending on the causal model, one can further choose between different Neyman-orthogonal score / moment functions. For the PLR model the default score is `"partialling out"`, i.e.,
\begin{align}\begin{aligned}\psi(W; \theta, \eta) &:= [Y - \ell(X) - \theta (D - m(X))] [D - m(X)].\end{aligned}\end{align}

The user guide provides details about the Sample-splitting, cross-fitting and repeated cross-fitting, the Double machine learning algorithms and the Score functions
Note that with this score, we do not estimate $g_0(X)$ directly, but the conditional expectation of $Y$ given $X$, $\ell_0(X) = E[Y|X]$. The user guide provides details about the Sample-splitting, cross-fitting and repeated cross-fitting, the Double machine learning algorithms and the Score functions


## Estimate double/debiased machine learning models
Expand All @@ -122,11 +123,11 @@ We now initialize `DoubleMLPLR` objects for our examples using default parameter

```{r}
set.seed(3141)
obj_dml_plr_bonus = DoubleMLPLR$new(dml_data_bonus, ml_g=ml_g_bonus, ml_m=ml_m_bonus)
obj_dml_plr_bonus = DoubleMLPLR$new(dml_data_bonus, ml_l = ml_l_bonus, ml_m = ml_m_bonus)
obj_dml_plr_bonus$fit()
print(obj_dml_plr_bonus)

obj_dml_plr_sim = DoubleMLPLR$new(dml_data_sim, ml_g=ml_g_sim, ml_m=ml_m_sim)
obj_dml_plr_sim = DoubleMLPLR$new(dml_data_sim, ml_l = ml_l_sim, ml_m = ml_m_sim)
obj_dml_plr_sim$fit()
print(obj_dml_plr_sim)
```
Expand Down