perform.impostors: An Authorship Verification Classifier Known as the Impostors...

In computationalstylistics/stylo: Stylometric Multivariate Analyses

An Authorship Verification Classifier Known as the Impostors Method. ATTENTION: this function is obsolete; refer to a new implementation, aka the imposters() function!

Description

A machine-learning supervised classifier tailored to assess authorship verification tasks. This function is an implementation of the 2nd order verification system known as the General Impostors framework (GI), and introduced by Koppel and Winter (2014). The current implementation tries to stick – as closely as possible – to the description provided by Kestemont et al. (2016: 88).

Usage

perform.impostors(candidate.set, impostors.set, iterations = 100,
               features = 50, impostors = 30,
               classes.candidate.set = NULL, classes.impostors.set = NULL,
               distance = "delta", z.scores.both.sets = TRUE) 

Arguments

candidate.set	a table containing frequencies/counts for several variables – e.g. most frequent words – across a number of texts written by a target author (i.e. the candidate to authorship). This table should also contain an anonymous sample to be assessed. Make sure that the rows contain samples, and the columns – variables (words, n-grams, or whatever needs to be analyzed).
impostors.set	a table containing frequencies/counts for the control set. This set should contain the samples by the impostors, or the authors that could not have written the anonymous sample in question. The variables used (i.e. columns) must match the columns of the candidate set.
iterations	the model is rafined in N iterations. A reasonable number of turns is a few dozen or so (see the argument "features" below).
features	the "impostors" method is sometimes referred to as a 2nd order authorship verification system, since it selects randomly, in N iterations, a given subset of features (words, n-grams, etc.) and performs a classification. This argument specifies the percentage of features to be randomly chosen; the default value is 50.
impostors	in each iteration, a specified number of texts from the control set is chosen (randomly). The default number is 30.
classes.candidate.set	a vector containing class identifiers for the authorial set. When missing, the row names of the set table will be used; the assumed classes are the strings of characters followed by the first underscore. Consider the following examples: c("Sterne_Tristram", "Sterne_Sentimental", "Fielding_Tom", ...), where the classes are the authors' names, and c("M_Joyce_Dubliners", "F_Woolf_Night_and_day", "M_Conrad_Lord_Jim", ...), where the classes are M(ale) and F(emale) according to authors' gender. Note that only the part up to the first underscore in the sample's name will be included in the class label.
classes.impostors.set	a vector containing class identifiers for the control set. When missing, the row names of the set table will be used (see above).
distance	a kernel (i.e. a distance measure) used for computing similarities between texts. Available options so far: "delta" (Burrow's Delta, default), "argamon" (Argamon's Linear Delta), "eder" (Eder's Delta), "simple" (Eder's Simple Distance), "canberra" (Canberra Distance), "manhattan" (Manhattan Distance), "euclidean" (Euclidean Distance), "cosine" (Cosine Distance). THIS OPTION WILL BE CHANGED IN NEXT VERSIONS.
z.scores.both.sets	many distance measures convert input variables into z-scores before computing any distances. Such a variable weighting is highly dependent on the number of input texts. One might choose either training set only to scale the variables, or the entire corpus (both sets). The latter is default. THIS OPTION WILL BE CHANGED (OR DELETED) IN NEXT VERSIONS.

Value

The function returns a single score indicating the probability that an anonymouns sample analyzed was/wasn't written by a candidate author. As a proportion, the score lies between 0 and 1 (higher scores indicate a higher attribution confidence).

Author(s)

Maciej Eder

References

Koppel, M. , and Winter, Y. (2014). Determining if two documents are written by the same author. "Journal of the Association for Information Science and Technology", 65(1): 178-187.

Kestemont, M., Stover, J., Koppel, M., Karsdorp, F. and Daelemans, W. (2016). Authenticating the writings of Julius Caesar. "Expert Systems With Applications", 63: 86-96.

See Also

imposters

computationalstylistics/stylo documentation built on April 7, 2024, 4:12 p.m.