Unsupervised Learning in Machine Learning

Instructor: Stephen Perkins

Stephen is a technology and electronics expert who has a passion for the work that he does.

Have you ever wondered how Netflix knows what show you might like to watch next? This lesson will take you through the definition of an unsupervised learning algorithm, how it works, and a real-world example.

Unsupervised Learning Machines

Unsupervised learning is just one of the many machine learning algorithms used by a machine in order to achieve artificial intelligence. This machine learning type groups similar information according to specific variables without the need for predetermined labels. This essentially allows a machine to teach itself without the need for any user input, unlike supervised learning. Many call this the true form of artificial intelligence since the computer can think and solve problems by itself. There are two ways we can group unsupervised learning, clustering, and association.

  • Clustering - This is the most common learning model that is used to help identify the hidden patterns or groups within a data set.
  • Association - A learning model for when you want to identify the relationship between variables using large amounts of information from your data set. Companies like Amazon, Netflix, and Hulu use this to predict what items you might want to purchase or shows you might like to watch based on your account history.

Are There Downsides?

Unsupervised learning is great because it allows a machine to perform complex tasks on its own without human interaction, but this also creates a new hurdle for the machine to overcome. Since there is no data set being given to the machine, the outcome generally becomes an unknown factor. This means it starts out knowing nothing and has to sort out the information using only the input data in order to make sense of the problem at hand. Occasionally, it might not be able to sort the information into groups correctly and may end up creating more clusters instead of a clean organization. This is one of the most notable potential outcomes to expect if the information from the data set has too many complex items or similarities that can't be sorted into precise groups.


You can sometimes end up with more clusters than organization.
You can sometimes end up with more clusters than organization.


The Most Common Clustering Algorithm

We know that clustering is a technique often used in unsupervised learning situations, and there is a commonly used clustering algorithm that we should be aware of. K-means clustering is one of the most straightforward algorithms used for unsupervised learning. The machine uses unlabeled data to find the groups from the data set using the variable 'K.' Some points have to be randomly generated or selected at the start called cluster centroids, which act as a means of grouping data points based on similarities. These said individual clusters are essentially averaged and moved into the centroid that it fits most accurately. It will continuously go through this process until there are no more changes in the data set; thus the groupings should be as accurate as possible.


This is what a perfect solution to a clustering problem looks like.
This is what a perfect solution to a clustering problem looks like.


A Real World Example

One of the best ways to help you visualize how a machine uses unsupervised learning is to look at a real-world example. Let's say you go to your local park and decide you want to try and divide all of the people you see into their very own centroids. You do not know any of these people personally, so you do not have any predetermined data about them. For example, you might decide to divide them up by age, ethnicity, gender, type of hair, or even the type of shoes they are wearing.

To unlock this lesson you must be a Study.com Member.
Create your account

Register for a free trial

Are you a student or a teacher?

Unlock Your Education

See for yourself why 30 million people use Study.com

Become a Study.com member and start learning now.
Become a Member  Back
What teachers are saying about Study.com
Free 5-day trial

Earning College Credit

Did you know… We have over 160 college courses that prepare you to earn credit by exam that is accepted by over 1,500 colleges and universities. You can test out of the first two years of college and save thousands off your degree. Anyone can earn credit-by-exam regardless of age or education level.

To learn more, visit our Earning Credit Page

Transferring credit to the school of your choice

Not sure what college you want to attend yet? Study.com has thousands of articles about every imaginable degree, area of study and career path that can help you find the school that's right for you.

Create an account to start this course today
Try it free for 5 days!
Create an account
Support