PolishedCode_MachineLearning

Coordinate Descent Algorithm with Elastic Net Regularization

This is a polished code example of my own implementation of a Coordinate Descent Algorithm with Elastic Net Regularization used for solving least-squares regression for the minimization problem seen below:

This package was created as a part of my DATA 558 Machine Learning course at the University of Washington.

For examples of implementation, please see the examples folder.

No data files are required for download, as they are downloaded into the notebook via URLs.

For viewing the raw Python code for this implementation, please refer to myelasticnet.py

Directory Structure

PolishedCode_MachineLearning/

  |- examples/
     |- __init__.py
     |- README.md
     |- Polished Code - ElasticNet (Comparison with Sci-Kit Learn).ipynb
     |- Polished Code - ElasticNet (Real-World Example).ipynb
     |- Polished Code - ElasticNet (Simulated Example).ipynb
  |- images/
     |- elasticnet.jpg
  |- src/
     |- __init__.py
     |- myelasticnet.py
  |- README.md
  |- setup.py

Installation

Note: To run this package you will need familiarity with bash command line and Jupyter Notebook.

In a directory on your local machine, run the following git command in the bash terminal to clone the PolishedCode_MachineLearning repository onto your computer:

git clone https://github.com/samirpdx/PolishedCode_MachineLearning.git

Then in the bash terminal go to the the newly installed package folder:

cd PolishedCode_MachineLearning/

And install the package by running the setup.py file:

python setup.py install