PIPELINE CLASSIFICATION

Table of contents

• How to cite?
• Contents overview
• Installing environment

How to cite?

Contents overview

src

This directory contains scripts and notebooks used to launch analyses.

• preprocess.py is used to preprocess data (resampling, masking, normalization).
• training.py is used to launch the training of the pipeline classfier.

To launch these, just change the parameters in the script files and launch in terminal: python3 {script_name}.py.

results

This directory contains a notebook in which you can launch the test of the classifier.

data

This directory contains csv files of the datasets used to train the models.

Installing environment

To reproduce the figures, you will need to create a conda environment with the necessary packages.

• First, download and install miniconda or miniforce if you work on a Mac Apple Silicon.
• After this, you will need to create you own environment by running :

conda create -n workEnv
conda activate workEnv

The name workEnv can be changed.

• When you environment is activated, just run the install_environment.sh script available at the root of this directory.

bash install_environment.sh

This script will install all the necessary packages to reproduce this analysis. At each step, you might have to answer y/n, answer yes any time to properly do the install.