Re: [R] help with LDA topic modelling..
Thanks a lot From: Jim Lemon Sent: Monday, December 20, 2021 1:43 PM To: akshay kulkarni Cc: R help Mailing list Subject: Re: [R] help with LDA topic modelling.. Hi Akshay, It depends upon how the circles are calculated. If each circle encloses all of the members of each group, yes. I doubt you would get perfect separation in any real example, though. At the moment: Hanukkah is finished, Margashirsha Punima has just happened and Christmas is coming up. No matter how devout the adherents, I don't think you would find news articles that only mentioned one of these festivals. Less facetious groupings would suffer the same eclectic problem. Jim On Mon, Dec 20, 2021 at 6:29 PM akshay kulkarni wrote: > > dear members, > I am using LDA for topic modelling of news articles > (topicmodels package). I am visualizing the accuracy with the LDAvis package. > > The visualization shows clusters as circles, probably intersecting. My > question is, if a find the optimal number of topics, k, and if the circles > representing the topics doesn't intersect, then I have achieved perfect > segregation. AM I right? > > Thanking You, > Yours sincerely, > AKSHAY M KULKARNI > > [[alternative HTML version deleted]] > > __ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] help with LDA topic modelling..
Hi Akshay, It depends upon how the circles are calculated. If each circle encloses all of the members of each group, yes. I doubt you would get perfect separation in any real example, though. At the moment: Hanukkah is finished, Margashirsha Punima has just happened and Christmas is coming up. No matter how devout the adherents, I don't think you would find news articles that only mentioned one of these festivals. Less facetious groupings would suffer the same eclectic problem. Jim On Mon, Dec 20, 2021 at 6:29 PM akshay kulkarni wrote: > > dear members, > I am using LDA for topic modelling of news articles > (topicmodels package). I am visualizing the accuracy with the LDAvis package. > > The visualization shows clusters as circles, probably intersecting. My > question is, if a find the optimal number of topics, k, and if the circles > representing the topics doesn't intersect, then I have achieved perfect > segregation. AM I right? > > Thanking You, > Yours sincerely, > AKSHAY M KULKARNI > > [[alternative HTML version deleted]] > > __ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] help with LDA topic modelling..
dear members, I am using LDA for topic modelling of news articles (topicmodels package). I am visualizing the accuracy with the LDAvis package. The visualization shows clusters as circles, probably intersecting. My question is, if a find the optimal number of topics, k, and if the circles representing the topics doesn't intersect, then I have achieved perfect segregation. AM I right? Thanking You, Yours sincerely, AKSHAY M KULKARNI [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
