Machine Learning from Scratch in Java

This repository contains Java implementations of core machine learning algorithms built from scratch without using external ML libraries.

The goal of this project is to develop a deeper understanding of how common machine learning models work under the hood through manual implementation.

Algorithms Implemented

• Linear Regression
• Logistic Regression
• MNIST Digit Classification

Key Concepts Covered

• Gradient Descent
• Loss Functions
• Model Training and Evaluation
• Numerical Optimization
• Data Processing in Java

Technologies Used

• Java
• Basic file I/O for data handling
• No external machine learning libraries

Project Context

This project was developed as part of an introductory machine learning course and focuses on understanding the mathematical and algorithmic foundations behind machine learning models rather than relying on pre-built frameworks.

How to Run

Each algorithm is implemented in its own Java file and can be compiled and run independently using a standard Java compiler.