Quantitative methods for MNIST: Linear Regression and SVM

This project demonstrates the implementation and evaluation of two fundamental machine learning models: Linear Regression and Support Vector Machine (SVM). It includes custom implementations for educational purposes, emphasizing how these algorithms function under the hood.

Project Structure

.
├── main.py                # Entry point for training and evaluating models
├── linear_regression.py   # Linear Regression model implementation
├── svm.py                 # Support Vector Machine implementation
├── utils.py               # Utility functions for data generation and evaluation

Features

• Custom implementation of Linear Regression using gradient descent
• Support Vector Machine with hinge loss
• Synthetic dataset generation for binary classification
• Visualization of model performance and decision boundaries (in progress)

Requirements

To install the required packages, run:

pip install numpy matplotlib

Usage

Run the main script to train and evaluate both models:

python main.py

This will:

• Generate synthetic datasets
• Train Linear Regression and SVM models
• Display evaluation metrics and decision boundaries (in progress)

Example Output

• Accuracy metrics for both models
• Plots showing the classification boundaries and data distribution (in progress)

License

This project is open-source and available under the MIT License.

---

Feel free to explore and modify the models for experimentation or educational use!