# Why do final cluster centers change after applying results from past K-Means clustering (SPSS)?

I have a question regarding what happens after I apply k-means clustering centers to a new data set.

Basically, I ran k-means clustering on a dataset1, saved the cluster centers, and applied it to a new dataset2 (set SPSS to "read initial" cluster centers and set the methodology to "classify only"). SPSS then outputs the clusters for my new dataset2. In the output however, there is also a set of initial and final cluster centers.

The initial cluster centers are the same ones that I've loaded in, but what is the final cluster centers shown here and why do they differ from the initial cluster centers that I loaded? Is my data still being classified using the cluster centers that I've loaded?

Thank you!

• It will be hard for anyone who doesn't have SPSS to tell you. In general, software-specific questions are off-topic here. I do think there may be a statistical question here, but the answer you get may not include any information about how SPSS works. – gung Apr 2 '15 at 1:58
• hi @gung, thanks for the note! i figured the answer would relate to both SPSS and k-means methodology -- as in, when you "reapply" cluster centers to a new data set, the cluster centers shouldn't shift, right? if so, why do the "final cluster centers" show up as different? – mikeyonaboat Apr 2 '15 at 4:21
• It may, but it also may not. Keep your fingers crossed & you may get both answers. I think there is enough statistical content for this to be on-topic. – gung Apr 2 '15 at 4:23
• Is my data still being classified using the cluster centers that I've loaded? Yes. The points were assigned to their nearest input centers. No further reassignment is done: the process stops. It displays you the recalculated centres after that assignment. These final centers were never used to reassign. – ttnphns Apr 2 '15 at 9:05