q_collocations: Collocations to query match

In wiertz/rusecqp: Corpus linguistic analysis for Corpus Workbench corpora

Description

q_collocations performs collocation analysis of a query match. If query match is longer than one token the context is build around the entire length of the query match. E.g. if context = 3, three tokens before the first token of the match and three tokens after the last token of the match are taken into account. In case two or more numeric collocation windows overlap, tokens are counted multiple times. However, if more than one match fall within a structural region, only unique structural regions are considered.

Usage

q_collocations(query_result, context, pattr = "word", exclude_match = T)

Arguments

query_result	list of class query result returned from query_corpus.
context	integer vector of length one: number of tokens around the query match. integer vector of length two: number of tokens to the left and right. character: structural attribute (e.g. "s").
pattr	character specifying positional attribute (e.g. "word").
exclude_match	logical indicating whether the matched tokens themselves should be excluded from the analysis.

Value

data.table of collocators and frequencies.

Examples

## Not run: q_collocations(my_query_result, context = 3)
  ## Not run: q_collocations(my_query_result, context = c(2, 0))
  ## Not run: q_collocations(my_query_result, context = "s", pattr = "lemma", exclude_match = F)

wiertz/rusecqp documentation built on Feb. 9, 2022, 1:30 p.m.