https://github.com/msikorski93/linear-regression-from-scratch

Different solutions of linear and polynomial regression made from scratch.
https://github.com/msikorski93/linear-regression-from-scratch

from-scratch linear-regression numpy polynomial-regression tensorflow

Last synced: 8 months ago
JSON representation

Different solutions of linear and polynomial regression made from scratch.

• Host: GitHub
• URL: https://github.com/msikorski93/linear-regression-from-scratch
• Owner: msikorski93
• Created: 2023-04-27T21:58:08.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-12-29T16:38:47.000Z (9 months ago)
• Last Synced: 2024-12-29T17:25:31.966Z (9 months ago)
• Topics: from-scratch, linear-regression, numpy, polynomial-regression, tensorflow
• Language: Jupyter Notebook
• Homepage:
• Size: 982 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Linear-Regression-From-Scratch

![ alt text ](https://img.shields.io/badge/license-MIT-green?style=&logo=)

![ alt text ](https://img.shields.io/badge/-Jupyter-F37626?logo=Jupyter&logoColor=white)

![ alt text ](https://img.shields.io/badge/-NumPy-013243?logo=Numpy&logoColor=white)

![ alt text ](https://img.shields.io/badge/-TensorFlow-FF6F00?logo=TensorFlow&logoColor=white)

![ alt text ](https://img.shields.io/badge/-Keras-D00000?logo=Keras&logoColor=white)

Linear and polynomial regression solutions made from scratch using TensorFlow 1 framework.

### New notebook for January 2025

Linear regression solutions built from scratch, without using fancy frameworks (only NumPy was implemented). We evaluated the performances with basic regression metrics as below:

|           Model           |    MSE   |    MAE   |    R²    |    RSS   |   Slope  | Intercept |

|:-------------------------:|:--------:|:--------:|:--------:|:--------:|:--------:|:---------:|

|      Gradient Descent     | 0.056114 | 0.208579 | 0.863454 | 6.733636 | 2.046441 | 11.358466 |

| Ordinary Least Squares #1 | 0.056114 | 0.20858  | 0.863454 | 6.733636 | 2.046379 | 11.358499 |

| Ordinary Least Squares #2 | 0.056114 | 0.20858  | 0.863454 | 6.733636 | 2.046379 | 11.358499 |

|        Scikit-Learn       | 0.056114 | 0.20858  | 0.863454 | 6.733636 | 2.046379 | 11.358499 |

We proved that our custom models are valid and accurate alternatives to the scikit-learn model, with virtually no noticeable difference in performance or output.