word_proximity: Proximity Matrix Between Words

Description

word_proximity - Generate proximity measures to ascertain a mean distance measure between word uses.

Usage

word_proximity(
  text.var,
  terms,
  grouping.var = NULL,
  parallel = TRUE,
  cores = parallel::detectCores()/2
)
# S3 method for word_proximity
weight(x, type = "scale", ...)

Value

Returns a list of matrices of proximity measures in the unit of average sentences between words (defaults to scaled).

Arguments

text.var: The text variable.
terms: A vector of quoted terms.
grouping.var: The grouping variables. Default NULL generates one word list for all text. Also takes a single grouping variable or a list of 1 or more grouping variables.
parallel: logical. If TRUE attempts to run the function on multiple cores. Note that this may not mean a speed boost if you have one core or if the data set is smaller as the cluster takes time to create.
cores: The number of cores to use if parallel = TRUE. Default is half the number of available cores.
x: An object to be weighted.
type: A weighting type of: c("scale_log", "scale", "rev_scale", "rev_scale_log", "log", "sqrt", "scale_sqrt", "rev_sqrt", "rev_scale_sqrt"). The weight type section name (i.e. A_B_C where A, B, and C are sections) determines what action will occur. log will use log, sqrt will use sqrt, scale will standardize the values. rev will multiply by -1 to give the inverse sign. This enables a comparison similar to correlations rather than distance.
...: ignored.

Details

Note that row names are the first word and column names are the second comparison word. The values for Word A compared to Word B will not be the same as Word B compared to Word A. This is because, unlike a true distance measure, word_proximity's matrix is asymmetrical. word_proximity computes the distance by taking each sentence position for Word A and comparing it to the nearest sentence location for Word B.

Examples

Run this code

if (FALSE) {
wrds <- word_list(pres_debates2012$dialogue, 
    stopwords = c("it's", "that's", Top200Words))
wrds2 <- tolower(sort(wrds$rfswl[[1]][, 1]))

(x <- with(pres_debates2012, word_proximity(dialogue, wrds2)))
plot(x)
plot(weight(x))
plot(weight(x, "rev_scale_log"))

(x2 <- with(pres_debates2012, word_proximity(dialogue, wrds2, person)))

## The spaces around `terms` are important
(x3 <- with(DATA, word_proximity(state, spaste(qcv(the, i)))))
(x4 <- with(DATA, word_proximity(state, qcv(the, i))))
}