As a human being, we learn from past experiences. We haven't become so charming by accident. Years of positive compliments as well as negative criticism have all helped shape us into who we are today. We learn how to ride a bike by trying out different muscle movements until it just clicks. When you perform actions, you're sometimes rewarded immediately. This is all about reinforcement learning (RL).
This lesson is all about designing a machine learning system driven by criticisms and rewards. We will see how to apply reinforcement learning algorithms for the predictive model on real-life datasets.
In a nutshell, the following topics will be covered throughout this lesson:
- Reinforcement learning
- Reinforcement learning for predictive analytics
- Notation, policy, and utility in RL
- Developing a multiarmed bandit's predictive model
- Developing a stock price predictive model
From a technical perspective, whereas supervised and unsupervised learning appears at opposite ends of the spectrum, RL exists somewhere in the middle. It's not supervised learning because the training data comes from the algorithm deciding between exploration and exploitation. And it's not unsupervised because the algorithm receives feedback from the environment. As long as you are in a situation where performing an action in a state produces a reward, you can use reinforcement learning to discover a good sequence of actions to take the maximum expected rewards.
The goal of an RL agent will be to maximize the total reward that it receives in the long run. The third main sub element is the value function.
While the rewards determine an immediate desirability of the states, the values indicate the long-term desirability of states, taking into account the states that may follow and the available rewards in these states. The value function is specified with respect to the chosen policy. During the learning phase, an agent tries actions that determine the states with the highest value, because these actions will get the best amount of reward in the long run.