Problem in the output of k-means

Discuss statistics related things
Post Reply
Posts: 8
Joined: Fri Feb 01, 2019 7:49 am

Problem in the output of k-means

Post by fabstat »

I realised a k-means analysis in Jamovi (Snowcluster package) and I have found a problem in the output.

The label of the clusters in the table Centroids of cluster Table is not the same than in the graph Plot of means across clusters. See the file attached:
Cluster 4 which is the maximum on the first variable, is labelled cluster 1 in the graph.

If I save the cluster membership and make descriptive statistics , I find the same results than in the Table.

Do I understand incorrectly the outputs ?

Thanks for your help.
Capture d’écran 2023-11-27 à 15.23.02.png (307.56 KiB) Viewed 2760 times
User avatar
Posts: 155
Joined: Mon Dec 04, 2017 8:40 am
Location: Seoul, Korea

Re: Problem in the output of k-means

Post by seol »

Hi fabstat

Are you using latest version of snowCluster module(ver.7.2.6)? you can check it in jamovi library. Here is a example data(iris data) in folder, which produce correct results.

Best Regards
이미지 003.png
이미지 003.png (110.99 KiB) Viewed 2739 times
Post Reply