This question was asked in 2014: How to perform document classification (i.e. extracting topics from text)?

At the time user @V.E. gave a great little tutorial on LDA, but today, there's still no satisfactory way of doing modern topic modeling.

Yes one could always use ExternalEvaluate to run LDA or LHTA in another language e.g. sklearn. But I really want to see something that is top-level and fast, like an adversarial topic modeling network for example.

Are there any recent talks on this, 3rd party libraries, or undocumented features to look for modern topic modeling approaches in Mathematica? Perhaps a dev can comment if this is yet planned in 12.4/13?

  • 1
    $\begingroup$ This is interesting but sort of open-ended. You might want to post some details of what you have in mind e.g. what would comprise a training set, what is the range of outcomes, etc. $\endgroup$ Aug 11, 2021 at 14:31


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.