0
$\begingroup$

I've applied K-Means clustering to my standardized data, including columns like age and salary. I have obtained the necessary coefficients. Now, when determining the cluster for a new data point, I currently use the actual age and salary values without standardization. Is this approach correct, or should I standardize any new data point before performing a sum product with the coefficients to identify the cluster?

Thanks, Anand

$\endgroup$
1
  • 1
    $\begingroup$ Hi @AnandRaj, welcome to the site. Can you describe in more detail what coefficients you are referring to and how you compute the "sum product with the coefficients to identify the cluster"? $\endgroup$
    – noe
    Commented Jan 11 at 12:20

1 Answer 1

1
$\begingroup$

K-Means learns clusters by looking at the distance between points. You want to make sure that any new point you want to assign to a cluster has been preprocessed the same way as your data. So if you standardized your data to find clusters, you should also standardize new data points.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.