Document Cluster Outputs in Text ExplorerI am hoping to find some output in Text Explorer > Latent Semantic Analysis, SVD > Cluster Documents, that is analogous to what you would get in the Hierarchical Cluster platform (e.g., the cluster history, and scree plot). Does anyone know a way to get this output in Text Explorer, or a way to recreate the results from Text Explorer in Hierarchical Cluster?
Re: Document Cluster Outputs in Text ExplorerHi Katie!
I am not sure if you can do that within the text explorer, but I'd recommend to save the document singular vectors to the data table and to use the hierarchical cluster analysis platform on the singular vectors. With that you'd have all options from the platform (including different options for clustering). Hope that helps!
Re: Document Cluster Outputs in Text ExplorerYes, I tried doing that, but was unable to reproduce the results in the Hierarchical Clustering Platform. I looked across the different distance metrics, but none were able to produce the same clusters (for an arbitrary k clusters). I'm not sure where I'm going astray... Any other thoughts appreciated.
Re: Document Cluster Outputs in Text Explorer@kaeng I talked with tech support this is how you recreate the Cluster variables.
To do this, save the Document Singular Vectors (red triangle option from the SVD).
Then put the singular vectors into Hierarchical Clustering, and use the Fast Ward method, without standardizing the data.
This should result in the same clustering as the Cluster Documents option.
