topfeatures {quanteda}

Identify the most frequent features in a dfm

Description

List the most (or least) frequently occurring features in a dfm, either as a whole or separated by document.

Usage

topfeatures(
  x,
  n = 10,
  decreasing = TRUE,
  scheme = c("count", "docfreq"),
  groups = NULL
)

Arguments

x

the object whose features will be returned

n

how many top features should be returned

decreasing

If TRUE, return the n most frequent features; otherwise return the n least frequent features

scheme

one of count for total feature frequency (within group if applicable), or docfreq for the document frequencies of features

groups

grouping variable for sampling, equal in length to the number of documents. This will be evaluated in the docvars data.frame, so that docvars may be referred to by name without quoting. This also changes previous behaviours for groups. See news(Version >= "3.0", package = "quanteda") for details.

Value

A named numeric vector of feature counts, where the names are the feature labels, or a list of these if groups is given.

Examples

dfmat1 <- corpus_subset(data_corpus_inaugural, Year > 1980) |>
    tokens(remove_punct = TRUE) |>
    dfm()
dfmat2 <- dfm_remove(dfmat1, stopwords("en"))

# most frequent features
topfeatures(dfmat1)
topfeatures(dfmat2)

# least frequent features
topfeatures(dfmat2, decreasing = FALSE)

# top features of individual documents
topfeatures(dfmat2, n = 5, groups = docnames(dfmat2))

# grouping by president last name
topfeatures(dfmat2, n = 5, groups = President)

# features by document frequencies
tail(topfeatures(dfmat1, scheme = "docfreq", n = 200))