pairwise_delta: Delta measure of pairs of documents

In widyr: Widen, Process, then Re-Tidy Data

Delta measure of pairs of documents

Description

Compute the delta distances (from its two variants) of all pairs of documents in a tidy table.

Usage

pairwise_delta(tbl, item, feature, value, method = "burrows", ...)

pairwise_delta_(tbl, item, feature, value, method = "burrows", ...)

Arguments

tbl	Table
item	Item to compare; will end up in item1 and item2 columns
feature	Column describing the feature that links one item to others
value	Value
method	Distance measure to be used; see dist()
...	Extra arguments passed on to squarely(), such as diag and upper

See Also

squarely()

Examples

library(janeaustenr)
library(dplyr)
library(tidytext)

# closest documents in terms of 1000 most frequent words
closest <- austen_books() %>%
  unnest_tokens(word, text) %>%
  count(book, word) %>%
  top_n(1000, n) %>%
  pairwise_delta(book, word, n, method = "burrows") %>%
  arrange(delta)

closest

closest %>%
  filter(item1 == "Pride & Prejudice")

# to remove duplicates, use upper = FALSE
closest <- austen_books() %>%
  unnest_tokens(word, text) %>%
  count(book, word) %>%
  top_n(1000, n) %>%
  pairwise_delta(book, word, n, method = "burrows", upper = FALSE) %>%
  arrange(delta)

# Can also use Argamon's Linear Delta
closest <- austen_books() %>%
  unnest_tokens(word, text) %>%
  count(book, word) %>%
  top_n(1000, n) %>%
  pairwise_delta(book, word, n, method = "argamon", upper = FALSE) %>%
  arrange(delta)

widyr documentation built on Sept. 13, 2022, 9:05 a.m.