Replication of Simulations in Bach et al. (forthcoming) - DoubleML - An Object-Oriented Implementation of Double Machine Learning in R

Philipp Bach Mrz 27, 2023

In this repository, we list and automatically run the simulation examples presented in Bach et al. (forthcoming)

Files for Replication of Figures and Results in DoubleML Package Vignette

The examples and results from the paper DoubleML - An Object-Oriented Implementation of Double Machine Learning in R can be reproduced with the R files listed in the following:

1. Section 4: Basic idea and key ingredients of double machine learning - Code for replication of simulation examples
   • examples_failure_n_500_p_20.R
2. Section 7.8: A short simulation study - Code for replication of simulation results
   • Cross-fitting: examples_failure_n_500_p_20.R
   • PLR: sim_plr.R
   • PLIV: sim_plivX.R
   • IRM: sim_irm.R
   • IIVM: sim_IIVM.R
   • Merging plots: merge_plots.R
   • Simultaneous inference: sim_siminf.R
3. Code chunks: The reproducible code contained in the code chunks is available via
   • doubleml_codechunks.R
   • Note: The code in doubleml_codechunks.R has been automatically produced from the manuscript (.Rmd) using knitr::purl(), see also Chapter 3.4 of Xi et al. (2020)

All data sets and DGPs used in the paper can be replicated via corresponding functions as provided in the DoubleML package. The required function calls are contained in the code chunks of the paper. The API documentation is available via https://docs.doubleml.org/r/stable/reference/index.html.

The current development version of DoubleML is available via the GitHub repository at https://github.com/DoubleML/doubleml-for-r. The stable version can be downloaded from CRAN https://cran.r-project.org/web/packages/DoubleML/index.html.

In case you have any questions, do not hesitate to contact philipp.bach@uni-hamburg.de

Citation

If you use the DoubleML package a citation is highly appreciated:

Bach, P., Chernozhukov, V., Kurz, M. S., Spindler, M., and Klaassen, S. (2021), DoubleML - An Object-Oriented Implementation of Double Machine Learning in R, arXiv:2103.09603.

@misc{DoubleML2021R,
  title={{DoubleML} -- {A}n Object-Oriented Implementation of Double Machine Learning in {R}},
  author={P. Bach and V. Chernozhukov and M. S. Kurz and M. Spindler and S. Klaassen},
  year={2021},
  eprint={2103.09603},
  archivePrefix={arXiv},
  primaryClass={stat.ML},
  note={arXiv:\href{https://arxiv.org/abs/2103.09603}{2103.09603} [stat.ML]}
}

Issues, Problems, Bugs

Please report problems and bugs as an issue in this repository.

Acknowledgements

We would like to thank Simon Couch for providing a detailed blogpost on how to use GitHub Actions for running R scripts.. This blogpost was used to set up the GitHub Actions used for automated replication of the simulation results.

References

Bach, P., Chernozhukov, V., Kurz, M. S., Spindler, M. and Klaassen, S. (2021), DoubleML - An Object-Oriented Implementation of Double Machine Learning in R, arXiv:2104.03220.

Xie, Y., Dervieux, C., & Riederer, E. (2020). R markdown cookbook. CRC Press.