Machine Learning From Scratch

Building fundamental machine learning algorithms from scratch in Python using only NumPy to understand how they work under the hood.

Purpose

This project focuses on educational implementations of core ML algorithms to build deep intuition about:

• Supervised Learning: Classification and regression fundamentals
• Distance-based methods: K-Nearest Neighbors, distance metrics
• Linear models: Linear regression, gradient descent optimization
• Neural Networks: Forward/backward propagation, activation functions
• Mathematical foundations: Loss functions, derivatives, matrix operations
• Training dynamics: Optimization algorithms, learning rates, convergence

The goal is to implement everything from first principles without relying on ML frameworks, gaining understanding of the underlying system.