BookmarkSubscribeRSS Feed
penguinflies90
Calcite | Level 5

I am doing k-means clustering using fastclus and am getting negative and decreasing CCC values as I increase the number of clusters. The following link says the distribution might be unimodal and long tailed.

 

http://support.sas.com/kb/22/540.html

 

What does this mean? Is this talking about the variables I am clustering on or the distribution of the CCC value itself? I am going to try non-parametric clustering to see if that gives me better results. Can anybody suggest anything else apart from that? 

 

Thanks

1 REPLY 1
MikeStockstill
SAS Employee

Hello Penguinflies90-

 

According to the link, "distribution" refers to the CCC.  A number-of-clusters solution is not clearly defined by the CCC in this case.  

Here is a note that provides some analysis strategies to consider.

 

46314: The Cluster node creates only one big cluster, or does not find clusters in the data

http://support.sas.com/kb/46/314.html

 

Have a great day.

SAS Innovate 2025: Save the Date

 SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!

Save the date!

How to choose a machine learning algorithm

Use this tutorial as a handy guide to weigh the pros and cons of these commonly used machine learning algorithms.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 1 reply
  • 1797 views
  • 0 likes
  • 2 in conversation