This repository hosts the R code implementation of the condition number regularized covariance estimator as an R package:

Won, J.-H., Lim, J., Kim, S.-J. and Rajaratnam, B. (2013), Condition-number-regularized covariance estimation. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 75: 427-450. doi:10.1111/j.1467-9868.2012.01049.x

install.packages("CondReg")

devtools::install_github("dddlab/CondReg")

Container-based development environments are also specified in this repository. The relevant files are:

• Files in .devcontainer directory
• docker-compose.yml file

Above files allow deploying the development environments using GitHub Codespaces, Visual Studio Code, or Docker compose.

Development environment can be deployed on your local machine (Dev Containers) or remotely on GitHub Codespaces. GitHub Codespaces can be accessed through a web-based Visual Studio Code interface. If you have a local installation of Visual Studio Code, it can connect to both Dev Containers and GitHub Codespaces.

You will need access to Codespaces and enabled. Then, the files in .devcontainer directory will do all the work!

Follow the Remote Development in Containers tutorial

Here are the steps:

# Clone this repository
git clone https://github.com/dddlab/CondReg
cd CondReg

# Build image and start container
docker-compose build CondReg
docker-compose up

Find and click on a link similar to

http://127.0.0.1:8888/lab?token=[generated token string]

and start your RStudio.

Due to the differences in working directory settings (workspaceFolder is not implemented in GitHub Codespaces), the path to the workspace folder is /workspaces/CondReg.

Since /workspaces directory is not writable by jovyan user, building and checking the R package would not work from /workspaces directory. Instead, there is a symbolic link directory /home/jovyan/work is also created. You would need to switch to /home/jovyan in order to build and check the R package:

cd /home/jovyan
R CMD build work
R CMD check CondReg_[version-number].tar.gz