Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/7307#issuecomment-120497662
About eta, you're right that specifying a Matrix that large would be silly
because of the size. One option would be to emulate
[http://nlp.stanford.edu/software/tmt/tmt-0.2/scaladocs/scaladocs/edu/stanford/nlp/tmt/lda/LDA.html],
which allows you to specify a Vector of length vocabSize (presumably setting
the same prior for each topic). Maybe it's not worthwhile though. Let's
create a separate JIRA and link to the Wallach paper + the Stanford NLP API,
and note that we're leaving it as a to-do for now to see if users have actual
needs for setting that prior.
Thanks for considering it!
Yes, I can leave the JIRA open after this gets merged.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
