Julia Silge has a great post on building Structural Topic Models in R using stm and tidytext:
The stm package has a
summary()method for trained topic models like these that will print out some details to your screen, but I want to get back to a tidy data frame so I can use dplyr and ggplot2 for data manipulation and data visualization. I can use
tidy()on the output of an stm model, and then I will get the probabilities that each word is generated from each topic.
I haven’t watched the video yet, but that’s on my to-do list for today.