Mean proportion of each topic within groups of documents,
e.g. by decade, journal or document type: a descriptive counterpart of
the regressions of topic_effects(), with no model assumption
and nothing smoothed.
Arguments
- theta
Document-topic matrix (
model$theta, documents in rows), or a fitted model.- meta
Metadata in the order of the rows of
theta.- by
Character vector of columns of
metadefining the groups.- weights
Optional numeric vector, one weight per document (e.g. its number of tokens), to weight documents in the average.
- ...
Not used: an argument renamed in the development version of mintR gives an error naming its replacement.
Value
A long data.table with the by columns, topic, prevalence
(weighted mean proportion; the topics of a group sum to 1) and
n_documents.
Examples
theta <- matrix(c(0.9, 0.1, 0.2, 0.8, 0.6, 0.4), ncol = 2, byrow = TRUE)
meta <- data.frame(period = c("a", "a", "b"))
topic_prevalence_by(theta, meta, by = "period")
#> period topic prevalence n_documents
#> <char> <int> <num> <int>
#> 1: a 1 0.55 2
#> 2: a 2 0.45 2
#> 3: b 1 0.60 1
#> 4: b 2 0.40 1