There are three types of Machine Learning, Supervised, Unsupervised and Reinforcement Learning. We will deal with the last two in future courses, but this course is all about Regression, which is a supervised learning approach.
Supervised Learning involves giving the machine learning algorithm the "right answer" in your data while you train it. If we go back to the example about setting the ideal price of a product, this means giving the algorithm the values of this price for previous products, so it can learn the behavior they contain.
On the other hand, unsupervised learning gives the machine learning algorithm only the features of the data, not the "right answer." Unsupervised learning is generally based around segmentation and clustering by similarity in features. For instance, we could use unsupervised learning to to analyze social media networks, clustering users together if they frequently interact with one another. We can also use it to recommend new movies to users given the previous movies they have already watched. As you can see with these examples, there isn't a "right answer" in the data, rather it's up to the machine learning model to learn patterns with the behavior of users.
Reinforcement Learning is a state-of-the-art technology, which involves the algorithm acting as a child does, learning based on the positive or negative reinforcement it is given. This has a lot of applications in creating AI to perform really well in video games like chess or snake.
Copyright © 2021 Code 4 Tomorrow. All rights reserved.
The code in this course is licensed under the MIT License.
If you would like to use content from any of our courses, you must obtain our explicit written permission and provide credit. Please contact classes@code4tomorrow.org for inquiries.