jacobeisenstein / sage Goto Github PK

Hi,
I read your paper, but do any instruction or documentation about how to start from scratch using the script here?

1. sample data to set an example?

Hi,

I saw the example you shared. However, I am not sure how to use your code for topic modeling(learning k topic from a collection of documents and infer topics for an unseen document using the trained model). Can you please provide an example?

Hi, first of all, thank you for researching this wonderful paper, SAGE!
While I'm reading your paper, I cannot clearly understand the background m (symboled as bold in the paper).
In paper, m is denoted by background distribution or background log-frequencies. In generative models, we draw it from uninformative prior.

What is m exactly and its meaning? Does it mean some basic values (or foundations) to calculate word probabilities ?? so that it does not matter if we draw it from uninformative prior such as uniform(0,1) ?

Thank you for reading this issue. Look forward to get your answer!