I write about the graph explained in this video that shows a big probability mass. It's a web-based interactive visualization of topics estimated using Latent Dirichlet Allocation that is built using a combination of R and D3; the default for scaling the set of inter-topic distances defaults to Principal Components, but other algorithms are also enabled. LDAvis

This is the way that it finally shows with some data I found online and used to understand Latent Dirichlet Allocation topic modeling. As far as I understand, this graph is ideal to give out final names to each topic according to the frequency or relevance of each term given the topic but I've got two questions:

1. What would you simply say about topic 8 position on the graph? Meaning that I can't tell what each square represents in this graph, I can't tell in plain words what's the difference between topic 1 and 10, or topic 8 and 3, given their position on this graph?

2. What can you say about the big gray line in comparison with the red filling when lambda is 0? I think I get that when lambda is 1, it's meaning about plain frequency of the term in the whole corpus vs the frequency of the terms in that topic, but I'm not sure I get it with putting more weight in the ratio from red to gray: what is the gray line saying about that term and how is red contrasting?

You can get this for a model named lda_model$plot() found here.

I am looking for the main idea at least. If there is info I need to add please tell me.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.