FFP17: Computational Studies of Adja Magatte Fall Internship

This repository contains supplementary material for the reproducibiliy of computational studies performed in the report Learning high dimensional gaussian graphical models written by Adja Magatte Fall under the supervision of:

• Pierre Fernique,
• Jean Peyhardi.

These studies are formatted as pre-executed Jupyter notebooks. Refers to the index.ipynb notebook which presents and references each study.

Test it !

Using Docker images and a Binder server, we are able to provide ways to reproduce the article studies without installing the StatisKit software suite.

Online with Binder

To reproduce the studies online, use this server.

On your computer with Docker

To reproduce the studies with Docker use these images. After installing Docker, you can type the following command in a shell:

docker run -i -t -p 8888:8888 statiskit/ffp17:latest

Then, follow the given instructions.

Install it !

You can also install required packages on your computer to reproduce these studies. In order to ease the installation of these packages on multiple operating systems, the Conda package and environment management system is used. For more information refers to the StatisKit software suite documentation concerning prerequisites to the installation step. Then, to install the required packages, proceed as as follows:

1. Clone this repository,

   git clone https://github.com/StatisKit/FFP17

2. Enter the cloned repository,

   cd FPD17

3. Install the given Conda environment,

   conda env create -f environment.yml

4. Activate the Conda environment as precised in your terminal.

5. Enter the share repository,

   cd share

6. Enter the jupyter repository,

   cd jupyter

7. Launch the Jupyter the index.ipynb notebook,

   jupyter notebook index.ipynb

8. Execute the index.ipynb notebook to execute all examples or navigate among referenced notebooks to execute them separatly.