bstewart/stm

Gamma.prior and Control parameters aren't changing anything

ricoblanc1 opened this issue · 0 comments

I want to adjust the number of topics per document (the equivalent of adjusting alpha in LDA), but for some reason I see no changes in my topic model output when I switch gamma.prior to L1 (from pooled) and/or when I try to adjust the gamma.enet through the control function. Is there something wrong with the way I've set up my code for this? No matter how I change those parameters, the topic model output I get ends up being the same.

Here's the final line of the R code I've been using (with various adjustments to what I put in the control list):

final <- stm(documents = out$documents, vocab = out$vocab,K = 50, prevalence =~ s(new.date), max.em.its = 50, gamma.prior = "L1", control=list(gamma.enet=.5), data = out$meta, init.type = "LDA", seed = 1)

summary (final)