Random correlation matrices with Metropolis-Hastings

This repository contains the files for replicating the experiments described in the paper

Córdoba I., Varando G., Bielza C., Larrañaga P. A fast Metropolis-Hastings method for generating random correlation matrices. Lecture Notes in Computer Science (IDEAL 2018), vol 11314, pp. 117-124, 2018.

Note that the time results described in the above paper may vary since we now use package randcorr for the polar parametrization, which is now computationally competitive with the rest of the compared methods.

Contents

• time_experiment.R: script that executes related state-of-the-art algorithms to measure execution time.
• plot_utils.R: utility functions for plotting.
• plot.R: script that generates the plots describing the properties of the algorithm, as well as the results of the time experiment.
• plot_comp.R: optional script for generating additional comparison results.

R packages required

These scripts rely on the R packages gmat, clusterGeneration and randcorr for executing the algorithms described in Córdoba et al. (2018). For plotting the graphics, it is also necessary to have installed the packages dplyr and ggplot2.

Usage

Rscript time_experiment.R
Rscript plot.R