How to interpret TAACO LDA indices?

[mayu...@sina.com]

Dear professors and peers,
               I have noticed that in the spreadsheet the LDA indices are "Average latent dirichlet allocation divergence score", that is, they are  divergence scores instead of similarity scores. So does it mean that the higher the LDA index value is, the lower the cohesion is? If so, do they apply to all LDA indices in TAACO? Thank you very much!

[Scott Crossley]

Higher LDA scores should relate to higher cohesion values. If ever curious, just run a correlation with LSA or word2vec scores. They should be positive, indicating higher cohesion. 

Let us know if that is not the case.

Scott 

On Thu, Jul 31, 2025 at 8:55 AM mayu...@sina.com <mayu...@sina.com> wrote:

Dear professors and peers,
               I have noticed that in the spreadsheet the LDA indices are "Average latent dirichlet allocation divergence score", that is, they are  divergence scores instead of similarity scores. So does it mean that the higher the LDA index value is, the lower the cohesion is? If so, do they apply to all LDA indices in TAACO? Thank you very much!

--
You received this message because you are subscribed to the Google Groups "Suite of automatic linguistic analysis tools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linguistic-analysi...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/linguistic-analysis-tools/f152ef64-c47b-4d5b-9fbd-7c16f4b3c5b2n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Scott Crossley

Professor, Department of Psychology and Human Development

Data Science Institute | Peabody College

Vanderbilt University

https://www.linguisticanalysistools.org/

https://learlab.org/