I love putting puzzles together, but the process can be awfully time consuming.
How long does it take you to put a puzzle together? What’s the most time consuming step for you?
For me, it’s grouping the pieces together. That seems to take me forever and is also rather boring. If I could, I would spend all of my time putting the puzzle pieces together. Not only is that more challenging, but it’s more fun!
What if you could push a button to group all the pieces together by a common trait like color? Let’s say your puzzle was a picture of a field and the sky. You could have all the blue pieces (probably sky) and green pieces (probably grass) grouped together for you then get right to putting them together.
Would this speed up the process of putting the puzzle together? What could you do with that extra time? How about put together more puzzles? 🙂
So why have I told you this story?
If you are following me here then you now know how the clustering machine learning algorithm works. It leverages unsupervised learning to evaluate data and put similar groups together without you even needing to tell it anything about the data. How cool is that?
Get our awesome product content delivered daily-ish to your inbox
But how would I apply this in the real world?
Well, VISA has used a similar technology to save $25B in fraudulent transactions in one year alone.
It works by searching for similar transactions and grouping them (just like the puzzle pieces). Then a person evaluates the patterns and focused on the fraudulent ones so you can take action. (put the puzzle together)
Accounting firms can leverage this technology to discovery fraud when conducting audits. This technology helps you do a more comprehensive search much faster. Take that fraud.
Fraud has been a massive, growing problem. This technology is our best weapon yet to eradicate it.
PS. You can read more about the k-means clustering algorithm here.