From f6fe3927b552e67fc8e97bd0320e5350957f4d8d Mon Sep 17 00:00:00 2001 From: Roland Stevenson Date: Mon, 11 Dec 2023 13:10:36 +0100 Subject: [PATCH 1/3] emphasize application to RCTs --- docs/about.rst | 8 +------- 1 file changed, 1 insertion(+), 7 deletions(-) diff --git a/docs/about.rst b/docs/about.rst index ccd5b09e..5c517f81 100644 --- a/docs/about.rst +++ b/docs/about.rst @@ -53,15 +53,9 @@ In essence, while traditional machine learning can tell us “what is”, causal Measuring Causal Effects ------------------------ -Different causal effects can be measured using varying techniques. - **Randomized Control Trials (RCT)** are the gold standard for causal effect measurements. Subjects are randomly exposed to a treatment and the Average Treatment Effect (ATE) is measured as the difference between the mean effects in the treatment and control groups. Random assignment removes the effect of any confounders on the treatment. -**Instrumental Variables (IV)** is a technique in which subjects are randomly exposed to a variable that influences treatment, but has no direct effect on the outcome. - -An example of an instrumental variable is a streamlined sign-up page that allows an Uber user to sign up for Uber Eats. Not all subjects will sign up for Uber Eats, but the streamlined sign-up makes it easier for them to experience both Uber and Uber Eats. Thereafter, the subject’s outcome is unrelated to their signup experience. In other words, the streamlined signup influences the treatment, not the outcome. - -If RCTs are not an option and hence confounders cannot be controlled for, the next best option is to attempt to **control for confounders** and measure the Conditional Average Treatment Effect (CATE). The CATE is an estimate of the treatment effect conditioned on all available covariates and confounders. Even if an RCT is available, if the treatment effects are heterogeneous across covariates, it might be preferable to measure the CATE. We call these Heterogeneous Treatment Effects (HTEs). +If an RCT is available and the treatment effects are heterogeneous across covariates, measuring the conditional average treatment effect(CATE) can be of interest. The CATE is an estimate of the treatment effect conditioned on all available experiment covariates and confounders. We call these Heterogeneous Treatment Effects (HTEs). Example Use Cases From 406fb3e3836f2c3b6b28a8c7f1182afa918c7843 Mon Sep 17 00:00:00 2001 From: Roland Stevenson Date: Mon, 11 Dec 2023 13:55:51 +0100 Subject: [PATCH 2/3] add Running Tests to installation.rst --- docs/installation.rst | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/docs/installation.rst b/docs/installation.rst index 3362477f..83c15ab8 100644 --- a/docs/installation.rst +++ b/docs/installation.rst @@ -102,3 +102,14 @@ Windows See content in https://github.com/uber/causalml/issues/678 +Running Tests +------------- + +Run all tests with: + +.. code-block:: bash + + pytest -vs tests/ --cov causalml/ + + +Add ``--runtf`` to run optional tensorflow tests which will be skipped by default. From 0aba84d553f283e0d6fe39522a823e1dd9ad6a93 Mon Sep 17 00:00:00 2001 From: Roland Stevenson Date: Mon, 11 Dec 2023 14:03:38 +0100 Subject: [PATCH 3/3] prioritize installation, quickstart, examples --- docs/index.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/index.rst b/docs/index.rst index 61966ea9..de64e353 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -7,10 +7,10 @@ Contents: :maxdepth: 2 about - methodology installation quickstart examples + methodology interpretation validation causalml