Code related to the D2C algorithm introduced in the JMLR paper "From Dependency to Causality: A Machine Learning Approach"

Details in http://jmlr.org/papers/v16/bontempi15a.html

To use it in your R code

• library(devtools); install_github("gbonte/D2C"); require(D2C)

• if (!requireNamespace("BiocManager", quietly = TRUE)) install.packages("BiocManager")

• BiocManager::install("graph")

• BiocManager::install("Rgraphviz")

• BiocManager::install("RBGL")

For sake of space the directory /data contains D2C models trained with small number of DAGs.

To change the directory to the one containing scripts

• setwd(find.package("D2C"))

To go to the directory of the scripts

• setwd(paste(find.package("D2C"),"scripts",sep="/"))

To list the D2C demos:

• dir(paste(find.package("D2C"),"scripts/",sep="/"),patt="demo*")

Pr. Gianluca Bontempi

co-Head of the Machine Learning Group

Département d'Informatique

Université Libre de Bruxelles

Boulevard du Triomphe - CP212

1050 Bruxelles, Belgium

email: [email protected]

Office Phone: +32-2-650 55 91

Fax: +32 2 650.56.09

http://mlg.ulb.ac.be