check_spelling: Check Spelling

In trinker/qdap: Bridging the Gap Between Qualitative Data and Quantitative Analysis

Description

check_spelling - Check the spelling for an vector of strings. The function use the following technique:

• Separate the words from a string into a bag of words.

• Look those words up in a dictionary to find words not recognized/found (considered possibly misspelled).

• These misses (possible misspellings) will be what is looked up for suggested replacements.

• Optionally, reduce dictionary by assuming the first letter of the misspelled word is correct (dictionary for this letter only).

• Reduce dictionary by eliminating words outside of the range of number of characters of the misspelled word.

• Use stringdist to find string distances between possible replacements and the misspelled term.

• Select n (n.suggests) terms from dictionary that are closest to the misspelled term.

which_misspelled - Check the spelling for a string.

check_spelling_interactive - Interactively check spelling.

correct - Access the spell corrector function from a "check_spelling_interactive" object for subsequent text character vector spelling corrections.

Usage

check_spelling(
  text.var,
  range = 2,
  assume.first.correct = TRUE,
  method = "jw",
  dictionary = qdapDictionaries::GradyAugmented,
  parallel = TRUE,
  cores = parallel::detectCores()/2,
  n.suggests = 8
)

which_misspelled(
  x,
  suggest = FALSE,
  range = 2,
  assume.first.correct = TRUE,
  dictionary = qdapDictionaries::GradyAugmented,
  method = "jw",
  nchar.dictionary = nchar(dictionary),
  first.char.dictionary = substring(dictionary, 1, 1),
  n.suggests = 8
)

check_spelling_interactive(
  text.var,
  range = 2,
  assume.first.correct = TRUE,
  click = TRUE,
  method = "jw",
  dictionary = qdapDictionaries::GradyAugmented,
  parallel = TRUE,
  cores = parallel::detectCores()/2,
  n.suggests = 8,
  ...
)

correct(x, ...)

Arguments

text.var	The text variable.
range	An integer of length 1 to use as a range for number of characters, beyond the number of characters of a word not found in the dictionary, to initially limit dictionary size and thus time to find a suggested replacement term. This may be expanded if no suitable suggestion is returned.
assume.first.correct	logical. If TRUE it is assumed that the first letter of the misspelled word is correct. This reduces the dictionary size, thus speeding up computation.
method	Method for distance calculation. The default is "jaccard". It is assumed that smaller measures indicate closer distance. Measures that do not adhere to this assumption will result in incorrect output (see stringdist for details).
dictionary	A character vector of terms to search for. To reduce overhead it is expected that this dictionary is lower case, unique terms.
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.
n.suggests	The number of terms to suggest. In the case of a tie (multiple terms have the same distance from misspelled word) all will be provided. Dictionary reduction may result in less than n.suggests suggested terms.
x	If which_misspelled - A character string. If correct - An object from check_spelling_interactive.
suggest	logical. If TRUE returns a data.frame with possible suggestions for misspelled words (words not found in the dictionary).
nchar.dictionary	A vector that corresponds in length and content to dictionary with elements that are the precalculated number of characters for each word in the dictionary.
first.char.dictionary	A vector that corresponds in length and content to dictionary with elements that are the pre-allotted first characters of each word in the dictionary.
click	logical. If TRUE the interface is a point and click GUI. If FALSE the interface is command line driven.
...	ignored

Value

check_spelling - Returns a data.frame with row (row number), not.found word.no (number of misspelled word), not.found (a word not found in the dictionary), suggestion (the most likely replacement for the word), and more.suggestions (A list of vectors of up to 10 most likely replacements).

which_misspelled - Returns either a named vector (names are the word number) of possible misspelled words (ifsuggestions = FALSE) or a data.frame with word.no (number of misspelled word), not.found (a word not found in the dictionary), suggestion (the most likely replacement for the word), and more.suggestions (A list of vectors of up to 10 most likely replacements).

check_spelling_interactive - Returns a character vector with the corrected text, the replacement list (via an attribute to the character vector), and a function to correct the same spelling errors in subsequent text character vectors.

correct - Returns a function for correcting spelling errors.

Note

A possible misspelled word is defined as not found in the dictionary.

check_spelling_interactive - The user may go back (undo) by pressing "TYPE MY OWN" entering either "!" (not) or "0" (similar to a phone system). The second choice in the "SELECT REPLACEMNT:" will be the original word and is prefixed with "IGNORE:". Press this to keep the original word.

References

https://stackoverflow.com/a/24454727/1000343
https://journal.r-project.org/archive/2011-2/RJournal_2011-2_Hornik+Murdoch.pdf

See Also

stringdist

Examples

## Not run: 
x <- "Robots are evl creatres and deserv exterimanitation."
which_misspelled(x, suggest=FALSE)
which_misspelled(x, suggest=TRUE)

check_spelling(DATA$state)

## browseURL("http://stackoverflow.com/a/24454727/1000343")
terms <- c("accounts", "account", "accounting", "acounting", "acount", "acounts", "accounnt")

set.seed(10)
(fake_text <- unlist(lapply(terms, function(x) {
    unbag(sample(c(x, sample(DICTIONARY[[1]], sample(1:5, 1)))))
})))

check_spelling(fake_text)

##============================##
## INTERACTIVE SPELL CHECKING ##
##============================##

## No misspellings found
check_spelling_interactive(DATA$state)

## character method approach (minimal example)
dat <- DATA$state; dat[1] <- "I likedd the cokie icekream"
(o <- check_spelling_interactive(dat))
preprocessed(o)
fixit <- attributes(o)$correct
fixit(dat)

## character method approach (larger example)
m <- check_spelling_interactive(mraja1spl$dialogue[1:75])
preprocessed(m)
fixit <- attributes(m)$correct
fixit(mraja1spl$dialogue[1:75])

## check_spelling method approach
out <- check_spelling(mraja1spl$dialogue[1:75])
(x <- check_spelling_interactive(out))
preprocessed(x)
correct(x)(mraja1spl$dialogue[1:75])
(y <- check_spelling_interactive(out, click=FALSE))
preprocessed(y)

## Examine Methods (?stringdist::stringdist)
strings <- c(
    "Robots are evl creatres and deserv exterimanitation kream.",
    "I gots me a biggert measrue, tommorrow"
)

meths <- c("osa", "lv", "dl", "hamming", "lcs", "qgram", "cosine", "jaccard", "jw")

stats::setNames(lapply(meths, function(x) check_spelling(strings, method=x)), meths)

## End(Not run)

trinker/qdap documentation built on Sept. 30, 2020, 6:28 p.m.