Regression Implementations

Implementations of linear, logistic, and softmax regression from scratch in Python using NumPy. This project demonstrates:

• Batch/stochastic gradient descent
• Multiclass classification
• Cross-entropy loss
• Forward/backward pass structure

Setup

Requires Python 3 with numpy and pytest.

pip install numpy pytest

Run Tests

Run a single file's tests:

pytest -v test_linear_regression.py
pytest -v test_logistic_regression.py
pytest -v test_softmax_regression.py

Or run everything at once:

pytest -v