The theta matrix is the predicted topic proportions.

multinom_theta(subsetted_dataset_item, ts_model, sim = 1)

Arguments

subsetted_dataset_item

list with (at least) element $full: unaltered dataset.

ts_model

TS fit to either the full dataset or a subset

sim

integer, which draw from the posterior to extract the thetas for

Value

matrix of predicted topic proportions (columns) for each timestep in the full dataset (rows)