Topic Modeling: LDA and BTM

Question

Does anyone know here about topic modeling? I badly need help.

1) What is Topic Modeling 2) What is Latent Dirichlet Allocation and Biterm Topic Modeling? 3) What is the difference between LDA and BTM? 4) How do they work?

I found studies but I can’t understand some of them especially BTM which is very few to find (even in YouTube, can’t find tutorial).

Thank you, so much.

score 1 · Answer 1 · answered Jan 17 '19 at 22:48

BTM is for clustering short text (e.g. survey answers, twitter data, short sentences), LDA is for clustering long text (e.g. news articles, whole papers). BTM clusters word-word cooccurrences, LDA clusters word-document occurrences. If you are into R,

BTM can be fit with R package BTM: https://cran.r-project.org/web//packages/BTM/index.html
LDA can be fit with R package topicmodels: https://cran.r-project.org/web//packages/topicmodels/index.html

Both links shown contain links to the math in papers

While the outcome of topic modelling can be seen as soft clustering, topic modelling's principal concern is inferring topics/themes from texts. — Rafs, Apr 23 '20 at 10:56

Topic Modeling: LDA and BTM

1 Answers1