Homemade Machine Learning: Python Implementations of Popular Machine Learning Algorithms with Jupyter Notebook Demos

This article introduces the Homemade Machine Learning GitHub project, which implements popular machine‑learning algorithms from scratch in Python, providing mathematical background, source code, and interactive Jupyter Notebook demonstrations for each algorithm.

Supervised learning covers regression (e.g., linear regression with single‑variable and multivariate examples) and classification (e.g., logistic regression with binary and multi‑class demos such as Iris flower classification and MNIST digit recognition). Each entry lists links to theory, code, and live demos.

Unsupervised learning includes clustering (K‑means with demos on Iris data) and anomaly detection (Gaussian‑based detection with examples like server‑operation outliers). Again, theory, implementation, and demo links are provided.

Neural networks are described as a framework rather than a single algorithm, with a focus on Multi‑Layer Perceptron (MLP) models and demos for MNIST digit and clothing classification.

The article also outlines the prerequisites for using the repository: install Python, then install all required packages with the command pip install -r requirements.txt , and launch Jupyter locally or remotely.

Data sets used by the project are available at the project's data folder . Several illustrative images are included to visualize concepts and project structure.

Additional resources such as Python performance optimization guides and curated GitHub repositories are linked at the end of the article.