Badr-MOUFAD/beyond-OLS-med-supp-material

Supplementary material for the medium article Beyond linear regression: Leveraging linear regression for feature selection of continuous/categorical variables.

Beyond linear regression

— Leveraging linear regression for feature selection of continuous/categorical variables—

Overview

This repository is a supplementary material for the medium article Beyond linear regression: Leveraging linear regression for feature selection of continuous/categorical variables.

It applies the introduced feature selection technic on the Automobile Data Set. The objective is to find the top $K$ most relevant features that explains the price of the car.

Project architecture

I adopt cookiecutter Simple DS project to structure this repository.

• data folder gathers raw and processed data
• notebooks contains the notebooks for preprocessing, exploring, and performing features selection.
• py_scripts is a python package where I put all the utils used in to produce the notebooks.

Get started

1. clone the repository on your local machine in cd to it

git clone https://github.com/Badr-MOUFAD/supp-material-med-article
cd supp-material-med-article

2. Initialize a the conda environnement and install py_scripts

conda env create -f environment.yml
pip install -e .

3. run the notebooks in the notebooks folder in the specified order

useful links

• Beyond linear regression medium article: https://towardsdatascience.com/beyond-linear-regression-467a7fc3bafb
• Automobile Dataset: https://archive.ics.uci.edu/ml/datasets/Automobile
• cookiecutter Simple DS project article: https://towardsdatascience.com/its-time-to-structure-your-data-science-project-1fa064fbe46
• celer documentation: https://mathurinm.github.io/celer/