sql server data mining:

I don't know if anyone in the community has, but it's a difficult task. Cluster characteristics isn't really the place to start. Why? Because all it tells you are the most probable contents in each cluster. If every single person in your dataset were Old, Fat, Austrailians, then every cluster's characteristics would be Old, Fat, and Austrailian.

A better thing to look at is a cluster discrimination vs. its complement, or the cluster description field from the content. These will tell you what's unique about the cluster, i.e. what differentiates is from the rest of the data.

[quoted text, click to view]