Techniques and tips for interpreting a cluster analysis

Question

Thanks to the posts and advice here, I've gotten off to a good start with cluster analysis. On a good day I can even get R to compute one without breaking down in tears! But I'm wondering if people have tips and techniques for interpreting what the clusters they get actually mean. By this I mean that when you sit down with a dendrogram or summary statistics from a cluster analysis, how to do you go about identifying what those clusters have in common? How do you apply meaning to the results?

I'd be interested in learning about any statistical or qualitative techniques people have found useful for analyzing clusters

CA is in essence an exploratory approach, so I'm afraid that whether cluster results make sense is a matter of your knowledge in the field of the study. However, if your question is about characterizing clusters properties afterwards, then there are a lot of possible way to describe the results using standard summary statistics. If so, it might be good to add a concrete example or reformulate your question in that direction. — chl, Jul 16 '11 at 20:55
Have you looked through questions/tagged/clustering+data-visualization, in particular Visualization-software-for-clustering ? — denis, Jul 20 '11 at 12:59

Techniques and tips for interpreting a cluster analysis

0 Answers0