I have never used Hierarchical Cluster Analysis for inferential statistics before, but the dendrogram it produces provides a nice way to visualise my data. I applied the HCA to my variables with the ClustOfVar R package. It broke down my variables into two separate clusters at the top of dendrogram.

Now, I know that my data conforms with the two-component hypothesis, since I've done Principal Component Analysis, and it indeed yields these two separate components (two-component structure is also theoretically sound).

I want to present the dendrogram simply for the visualisation, but feel uncomfortable doing so without any validation statistics. Are there any basic validation criteria for HCA that do not require robust assumptions, since I'm not using the actual clusters for inference?

Or would it be simply wrong to present an HCA dendrogram without actually using HCA for analysis?

Thank you.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.