Summary

In this chapter, you were first introduced to the underlying foundations of reinforcement learning. You saw that reinforcement learning models are focused on taking actions rather than on making predictions.

You also saw two widely known algorithms for reinforcement learning. This started with Q-learning, which is the foundational algorithm of reinforcement learning, and its more powerful improvement, Deep Q-learning.

Reinforcement learning is often used for more advanced use cases such as chatbots or self-driving cars, but can also be used for numerical data streams very well. Through a use case, you saw how to apply reinforcement learning to streaming data for finance.

With this chapter, you have come to the end of discovering the most relevant machine learning models for online learning. In the coming chapters, you will discover a number of additional tools that you will need to take into account in online learning and that have no real counterpart in traditional...