I have a corpus that contains different documents and also a date for each document (the metadata). After having calculated an STM-model, I would like to calculate the gamma values (the probabilities that each document is generated from each topic) per document, but subsetted by the specific date in order to see how the presence of topics in my corpus varies over time.
To put it short: I want to subset the STM by the metadata.
However, I can't get it to work. There is no information in the stm-documentation or the stm-vignette about this, but I know that is has to work somehow.