https://github.com/fratorgano/ridgeregressionproject

Ridge Regression project for the Statistical Methods for Machine Learning course
https://github.com/fratorgano/ridgeregressionproject

Last synced: 5 days ago
JSON representation

Ridge Regression project for the Statistical Methods for Machine Learning course

• Host: GitHub
• URL: https://github.com/fratorgano/ridgeregressionproject
• Owner: fratorgano
• Created: 2023-08-22T15:16:03.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-10-18T08:49:48.000Z (about 1 year ago)
• Last Synced: 2023-10-18T09:45:39.032Z (about 1 year ago)
• Language: Jupyter Notebook
• Size: 10.2 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# RidgeRegressionProject

This is the repository of the Ridge Regression project for the Statistical Methods for Machine Learning course

The report is in the file [Report.pdf](https://github.com/fratorgano/RidgeRegressionProject/blob/main/Report.pdf)

The file [ridgeReg.py](https://github.com/fratorgano/RidgeRegressionProject/blob/main/ridgeReg.py) contains the implementation of the Ridge Regression predictor

Main experiments:

* [ridge_numerical.ipynb](https://github.com/fratorgano/RidgeRegressionProject/blob/main/ridge_numerical.ipynb) - train and test the ridge regressor using only the numerical part of the dataset

* [Encoding.ipynb](https://github.com/fratorgano/RidgeRegressionProject/blob/main/Encoding.ipynb) - analyze the possible different encoding types and check performance

* [Encoding_alternative.ipynb](https://github.com/fratorgano/RidgeRegressionProject/blob/main/Encoding_alternative.ipynb) - analyze the possible different encoding types and check performance (with a different dataset preprocessing)

* [ridge_all.ipynb](https://github.com/fratorgano/RidgeRegressionProject/blob/main/ridge_all.ipynb) - train and test the ridge regressor using the whole dataset and try to find best value for parameters

Reproducibility: 

* All the data and graphs reported were generated using Python 3.10.12 on Ubuntu 22.04 using WSL2 on Windows 11

* All the python packages used, along with their versions, are available in the file [requirements.txt](https://github.com/fratorgano/RidgeRegressionProject/blob/main/requirements.txt)

* All the random seeds, where needed, were set to 1 to ensure reproducibility of the results