K-means: When you don't know what you're looking for
Which customers behave alike? Which products are used in similar ways? K-means clustering helps answer these types of questions by grouping data into small numbers of meaningful clusters based on similarity.
The “k” simply means how many groups you want to explore. You might ask “Are there 3 types of customers? Or 5?” K-means finds the mean average of each group and assigns items (data points) to the closest one, repeating the process until clear patterns appear.
It’s useful when your data is large and complex. You can use it when you don’t know what you’re looking for, and it automatically turns complex data into insights.
It’s widely used for things like customer segmentation or product recommendation engines.
4 ways to decide: AI or automation?
When teams aren’t clear on what they actually need, AI often becomes the default request because it’s seen as the modern solution. KNIME’s Iris Adae gives tips on when to use AI versus automation and why. Read more here.
Worth watching
I really like this series on Machine Learning Algorithms Explained, where ML algorithms are broken down and explained in the easiest and most practical way possible. He also has one on k-means. Watch here.