This repository is part of the supplementary material for the paper:
“Distinguishing Cause from Effect Using Quantiles: Bivariate Quantile Causal Discovery”
The file R/bqcd.R contains an implementation of our causal discovery method.
EXAMPLES
———————————————————
See R/benchmark-xyz for examples.
Benchmarks can be run from root directory.
CONTENTS
———————————————————
.
|-- R
| |-- baselines
| | |-- GRAN_EMD
| | `-- Slope
| | |-- Slope.R
| | |-- resit
| |-- baselines.R
| |-- benchmark-ANLSMN.R
| |-- benchmark-SIM.R
| |-- benchmark-Tuebingen.R
| |-- generate-ANLSMN-data.R
| |-- bqcd.R
| `-- utils
| |-- data_generators.R
| `-- read_tueb.R
|-- README
|-- code.Rproj
|-- data
| |-- ANLSMN_pairs
| | |-- AN
| | |-- AN-s
| | |-- LS
| | |-- LS-s
| | `-- MN-U
| |-- SIM_pairs
| | |-- SIM
| | |-- SIM-G
| | |-- SIM-c
| | `-- SIM-ln
| `-- tuebingen_benchmark
`-- results
PACKAGES REQUIRED
———————————————————
install.packages(c("rvinecopulib", "statmod", "qrnn", "quantregForest")) # for bQCD
install.packages("MASS") # for the data generators
install.packages(c("CAM", "Hmisc", "np", "energy", "data.table",
"boot", "energy", "gptk", "clue")) # for the baselines