Implementation of the data-compatible Fast Causal Inference (dcFCI) algorithm by Ribeiro & Heider, 2025, for robust causal discovery under latent confounding, unfaithfulness, and mixed data.
If you use this work, please cite:
Ribeiro, A. H., & Heider, D. (2025). dcFCI: Robust Causal Discovery Under Latent Confounding, Unfaithfulness, and Mixed Data. ArXiv preprint arXiv:2505.06542. https://arxiv.org/abs/2505.06542
@misc{ribeiro2025dcfcirobustcausaldiscovery,
title={dcFCI: Robust Causal Discovery Under Latent Confounding, Unfaithfulness, and Mixed Data},
author={Adèle H. Ribeiro and Dominik Heider},
year={2025},
eprint={2505.06542},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2505.06542},
}
This guide walks you through all necessary steps to install dependencies and set up the package environment in R, including how to install archived versions of required packages.
install.packages("BiocManager")
BiocManager::install(c("RBGL", "graph", "Rgraphviz"))package_list <- c(
"lmtest", "pscl", "brms", "dagitty", "ggm", "igraph",
"pcalg", "SEMgraph", "doFuture", "DOT", "jsonlite", "rsvg"
)
install.packages(package_list, dependencies = TRUE, repos = "http://cran.us.r-project.org")
# Install any missing packages
new_packages <- package_list[!(package_list %in% installed.packages()[,"Package"])]
if (length(new_packages)) {
install.packages(new_packages, dependencies = TRUE, repos = "http://cran.us.r-project.org")
}wget https://cran.r-project.org/src/contrib/Archive/BFF/BFF_3.0.1.tar.gzinstall.packages(c("BSDA", "hypergeo", "gsl"), dependencies = TRUE)
install.packages("./BFF_3.0.1.tar.gz", repos = NULL, type = "source")wget https://cran.r-project.org/src/contrib/Archive/MXM/MXM_1.5.5.tar.gzmxm_packages <- c(
"lme4", "doParallel", "relations", "Rfast", "visNetwork",
"energy", "geepack", "bigmemory", "coxme", "Rfast2", "Hmisc"
)
mxm_packages <- mxm_packages[!(mxm_packages %in% installed.packages()[,"Package"])]
if (length(mxm_packages)) {
install.packages(mxm_packages, dependencies = TRUE, repos = "http://cran.us.r-project.org")
}
install.packages("./MXM_1.5.5.tar.gz", repos = NULL, type = "source")The FCI.Utils R package is available at https://github.com/adele/FCI.Utils/
You can install the development version directly from GitHub:
install.packages("devtools", dependencies=TRUE)
devtools::install_github("adele/FCI.Utils", dependencies=TRUE)The following system libraries may be required for full functionality (e.g., for graphics rendering, parallel processing, or Rcpp-based packages). You can install them using:
sudo apt-get update
sudo apt-get install -y \
libmagick++-dev \
cargo \
librsvg2-dev \
libavfilter-dev \
libharfbuzz-dev \
libcurl4-openssl-dev \
libudunits2-dev \
cmake \
libsodium-dev \
libssl-dev \
libxml2-dev \
libgdal-dev \
libfontconfig1-dev \
libcairo2-dev
⚠️ These packages are typically needed for rendering graphs, working with web graphics, compiling C++ code, or handling advanced statistical routines.
- Ensure that your R version is up to date (≥ 4.2 recommended).
- If you're running on Windows or macOS, equivalent system tools may be needed (e.g., Rtools, Xcode).
- For questions or bug reports, please open an issue on the dcFCI GitHub repository.
