Hi, I am a newbie in pyro and trying to use my dataset for LDA model , so not by generating the data as in pyro example, but using my own dataset, which looks somewhat like:
where there is topic column which is just id of each topic, theme1 and theme2 - each topic has ids of 2 themes, keywords1 - counts of keywords in theme1 and keywords2 - number of keywords from theme1 appeared in theme2, A and B are weights.
So not all the keywords2 are presented and have to be calculated from keywords1 distribution over theme2, according to weights.
How would I use my data to run it with lda model?