online-ml/awesome-online-machine-learning

This project is a curated list, often called an "awesome list," that gathers high-quality resources about online machine learning—a branch of machine learning where models learn from data that arrives one piece at a time, rather than all at once. The list is maintained by the community and follows the popular "awesome" format.

The README starts with a clear definition: online machine learning updates models incrementally with each new data point, unlike traditional batch learning which processes whole datasets in one go.

What’s inside?

• Courses and books — Links to free lecture notes, textbooks, and online courses. Examples include "Machine Learning for Streaming Data with Python" and a course from MIT called "Online Methods in Machine Learning." Many focus on the practical side, like using the Vowpal Wabbit library.

• Blog posts — Short articles that explain concepts and offer tutorials. Topics range from real-time ML challenges to anomaly detection using the River library. Some are in English, others in Portuguese.

• Software — Split into two parts:

  • Modelling — Libraries and tools that implement online learning algorithms. Examples include River (Python), MOA (Java), and Vowpal Wabbit (C++). The list also notes that scikit-learn has some incremental capabilities, though not fully online.
  • Deployment — Tools to put online models into production, such as KappaML and a Django plugin called django-river-ml.

• Papers — Academic research papers organised into many sections: linear models, support vector machines, neural networks, decision trees, unsupervised learning, time series, drift detection, anomaly detection, metric learning, graph theory, ensemble models, expert learning, active learning, and more. Each paper has a direct link and a short note where relevant, like pointing to source code or an implementation.

How is it organised?

Everything sits in a clean list with clear headings, making it easy to jump to what you need. The list is not a tutorial—it’s a starting point. It assumes you already know a bit about machine learning but want to find the best material on the online variant.

Who is it for?

Students, researchers, and practitioners who want to learn about or work with streaming data. If you’re building systems that need to update predictions in real time, this list points you to the right techniques and tools.

A note on faithfulness: All the details above come directly from the README of this project. I haven’t added any extra features or facts beyond what the README states.