pyldavis "individual terms...interpreting..current topic" v. gensim's print_topics

[husseinkohy]

Hi,

I have noticed that the "individual terms" shown on the horizontal barchart of pyldavis prints different words from gensim's print_topics function (https://radimrehurek.com/gensim/models/ldamodel.html). in some cases I do not see one matching word between the two calls. changing lambda does not make a difference. any insight would be greatly appreciated. I will post the same question to gensim's mailing list. thanks

update: it seems that number of topic assigned to a text could be different. is there a way to match the topic number assigned by these two models?

[MCin22]

Hi there! I just started to work using gensim and pyLDAvis and also noticed similar problem. Is there any way to match topics between gensim output and pyLDAvis visualization?

I wanted to manually name generated topics based on the most relevant terms and to give labels to documents that are included to corpora, but because of mismatching between outputs from gensim's print_topics function and pyLDAvis' visualization I can't do it. (I checked out whether the most relevant terms occurs in documents that belong to specific topic and got nothing)

[shah-p]

Hi there! I have observed the same issue. It's been a while since you have raised this issue.
Did you find any answer for this? It's not even been discussed widely. So, it's hard to find the answer.

Any help would be much appreciated. Thanks.

[prachimehta2705]

Hi,
I have been facing the same issue as well and have been stuck for a while now. Please share any possible solution.
Thanks

[husseinkohy]

Hi there, i have not worked on this issue since. i do not remember the issue. sorry about that.

…

On Tue, Jul 2, 2019 at 12:42 AM prachimehta2705 ***@***.***> wrote: Hi, I have been facing the same issue as well and have been stuck for a while now. Please share any possible solution. Thanks — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#79?email_source=notifications&email_token=ABZENVDD7GUNYTLZ6WNUGRDP5LMFBA5CNFSM4CZLS7E2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZABGZI#issuecomment-507515749>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABZENVA3ITW7YRWR27JHDGLP5LMFBANCNFSM4CZLS7EQ> .

[shah-p]

Hi,
I have been facing the same issue as well and have been stuck for a while now. Please share any possible solution.
Thanks

Hi,
I noticed that the topic numbers in pyLDAvis and LDA output are mismatching but results are matching. for example, words of topic 0 of LDA output (from sci-kit in my case) are in topic 2 in pyLDAvis and so on for all other topics.
Check this and let me know if you are getting the same way.