Same cluster, new data. How to keep things stable.


I did a cluster of 100 clients and I would like to know how to asign a cluster to new clients without changing the centroids and distances. I would like to keep the clusters design fixed, but asign a cluster to newones without changing centroids and distances. My practical problem is that I dont want to have a diffent cluster description (number, distance, threshold and centroits) every month....any ideas?


You might want to check out the Cluster Assigner node which consumes the generated model and applied it new / un-clustered data. Hope this helps together with your clustering node?