/hte3

Causal machine learning pipeline using tlverse/sl3

Primary LanguageR

hte3 Package

This package is in development. Improved documentation and examples coming soon.For questions, bugs, or feature additions, feel free to contact me at 'lvdlaan@uw.edu'!

hte3: Causal Machine Learning of Heterogeneous Treatment Effects using sl3

The hte3 package equips users with tools for supervised causal machine learning of heterogeneous treatment effects, leveraging the sl3 package.

Key Features

  • Customizable Meta-learners of HTEs: Any supervised machine learning algorithm supported by the sl3 package can be turned into a meta-learner for heterogeneous treatment effects, including the DR-learner, R-learner, T-learner, and EP-learner of the CATE. For details on the usage of the sl3 R package, we refer to https://github.com/tlverse/sl3.

  • Novel Meta-learners of the CRR: Implements novel EP-learners of the log conditional relative risk (CRR).

For comprehensive information, consult the package documentation.

Installation

To install the hte3 package, use the following command:

if(!require(devtools)) {
  install.packages("devtools")
}
devtools::install_github("Larsvanderlaan/hte3")