1
$\begingroup$

I have three variables which I have graphed as a scatter plot. One for x, one for y, and one for whether the point is red or blue. There are clearly two clusters - one red cluster, and one blue cluster.

How do I identify the two clusters, and then set a tolerance and figure out what the rough boundary is between the two clusters (i.e. say with a tolerance of 5, any points where 50 < x < 55 and 60 < y < 65 would be considered in the boundary)? I know I could use K-means for cluster identification. But I'm not sure how to deal with the boundary problem.

$\endgroup$
1
$\begingroup$

Don't look for clustering - you already have your classes.

What you are looking for is a classificator.

More precisely, you want a understandable model.

Consider a decision tree, or a SVM.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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