oppose: Contrastive analysis of texts

Description Usage Arguments Details Value Author(s) References See Also Examples

View source: R/oppose.R

Description

Function that performs a contrastive analysis between two given sets of texts. It generates a list of words significantly preferred by a tested author (or, a collection of authors), and another list containing the words significantly avoided by the former when compared to another set of texts. Some visualizations are available.

Usage

1
2
3
4
5
6
7
oppose(gui = TRUE, path = NULL, 
         primary.corpus = NULL,
         secondary.corpus = NULL,
         test.corpus = NULL,
         primary.corpus.dir = "primary_set",
         secondary.corpus.dir = "secondary_set",
         test.corpus.dir = "test_set", ...)

Arguments

gui

an optional argument; if switched on, a simple yet effective graphical interface (GUI) will appear. Default value is TRUE.

path

if not specified, the current working directory will be used for input/output procedures (reading files, outputting the results, etc.).

primary.corpus.dir

the subdirectory (within the current working directory) that contains one or more texts to be compared to a comparison corpus. These texts can e.g. be the oeuvre by author A (to be compared to the oeuvre of another author B) or a collection of texts by female authors (to be contrasted with texts by male authors). If not specified, the default subdirectory primary_set will be used.

secondary.corpus.dir

the subdirectory (within the current working directory) that contains a comparison corpus: a pool of texts to be contrasted with texts from the primary.corpus. If not specified, the default subdirectory secondary_set will be used.

test.corpus.dir

the subdirectory (within the current working directory) that contains texts to verify the discriminatory strength of the features extracted from the primary.set and secondary.sets. Ideally, the test.corpus.dir should contain texts known to belong to both classes (e.g. texts written by female and male authors in the case of a gender-oriented study). If not specified, the default subdirectory test_set will be used. If the default subdirectory does not exist or does not contain any texts, the validation test will not be performed.

primary.corpus

another option is to pass a pre-processed corpus as an argument (here: the primary set). It is assumed that this object is a list, each element of which is a vector containing one tokenized sample. Refer to help(load.corpus.and.parse) to get some hints how to prepare such a corpus.

secondary.corpus

if primary.corpus is used, then you should also prepare a similar R object containing the secondary set.

test.corpus

if you decide to use test corpus, you can pass it as a pre-processed R object using this argument.

...

any variable produced by stylo.default.settings can be set here, in order to overwrite the default values.

Details

This function performs a contrastive analysis between two given sets of texts, using Burrows's Zeta (2007) in its different flavors, including Craig's extensions (Craig and Kinney, 2009). Also, the Whitney-Wilcoxon procedure as introduced by Kilgariff (2001) is available. The function generates a vector of words significantly preferred by a tested author, and another vector containing the words significantly avoided.

Value

The function returns an object of the class stylo.results: a list of variables, including a list of words significantly preferred in the primary set, words significantly avoided (or, preferred in the secondary set), and possibly some other results, if applicable.

Author(s)

Maciej Eder, Mike Kestemont

References

Eder, M., Rybicki, J. and Kestemont, M. (2016). Stylometry with R: a package for computational text analysis. "R Journal", 8(1): 107-21.

Burrows, J. F. (2007). All the way through: testing for authorship in different frequency strata. "Literary and Linguistic Computing", 22(1): 27-48.

Craig, H. and Kinney, A. F., eds. (2009). Shakespeare, Computers, and the Mystery of Authorship. Cambridge: Cambridge University Press.

Hoover, D. (2010). Teasing out authorship and style with t-tests and Zeta. In: "Digital Humanities 2010: Conference Abstracts". King's College London, pp. 168-170.

Kilgariff A. (2001). Comparing Corpora. "International Journal of Corpus Linguistics" 6(1): 1-37.

See Also

stylo, classify, rolling.classify

Examples

1
2
3
4
5
6
7
8
9
## Not run: 
# standard usage:
oppose()

# batch mode, custom name of corpus directories:
oppose(gui = FALSE, primary.corpus.dir = "ShakespeareCanon",
       secondary.corpus.dir = "MarloweSamples")

## End(Not run)

stylo documentation built on Oct. 9, 2018, 1:04 a.m.

Related to oppose in stylo...