The following is the output of the cluster centers I got from a cluster model (kmeans - 6 clusters)

3.371069, 3.920354, 3.629747, 3.700000, 3.988506, 3.740385

However, after segmenting the data into the 6 clusters and taking the average of the data for each of the 6 clusters, I get a different set of numbers than the above.

3.7, 3.6, 3.6, 3.8, 4.2, 3.3

My question is are cluster centers different than means?

Why am I getting different values?


1 Answer 1


They are the same.

When you run K-Means, the cluster center changes every iteration. In each iteration, cluster center or mean is given for that specific cluster. In the next iteration, a new cluster may be formed. The cluster centers need to be calculated again.

When the algorithm converges, you get the absolute means of the clusters formed. For most intents and purposes, they are the same.

  • 1
    $\begingroup$ Thanks Abhik, Why do you suppose I get different values when I calculate the means after segmenting the data? $\endgroup$
    – kiva
    Commented Mar 19, 2019 at 5:45
  • $\begingroup$ "by segmenting the data"? I assume you are talking about calculating the mean after the algorithm converges. You should not get such results since the prerequisite for convergence is very negligible change in mean values in consecutive iterations. $\endgroup$ Commented Apr 3, 2019 at 17:38

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.