Authorship Attribution: A Comparative Study of Three Text Corpora and Three Languages

Abstract

The first objective of this paper is carry out three experiments intended to evaluate authorship attribution methods based on three test-collections available in three different languages (English, French, and German). In the first we represent and categorize 52 text excerpts written by nine authors and taken from 19th century English novels. In the second we work with 44 segments from French novels written by eleven authors, mostly from the 19th century. In the third we extract 59 German text excerpts from novels published mainly during the 19th and the beginning of the 20th century, written by 15 authors. The second objective is to analyse performance differences obtained when using word types or lemmas as text representations, and the third objective is to evaluate three authorship attribution schemes, the first of which uses principal component analysis (PCA), the second applies the Delta approach, and the third corresponds to a new authorship attribution method based on specific vocabulary. This concept is computed for a given text (or author profile) and then compared with the entire corpus. Based on this information, we show how a distance measure can be derived and by means of the nearest neighbor approach we suggest a simple and efficient authorship attribution scheme. Based on three test collections and using either word types or lemmas as features, we demonstrate that the suggested classification scheme performs better than the PCA method, and slightly better than the Delta approach.

Comparative evaluation of term selection functions for authorship attribution

“…On the other hand, we can apply an automatic part-ofspeech tagger. However, such an approach is not error-free, and some recent studies have compared the relative merits of these two text representation schemes for authorship attribution (Savoy, 2012;Miranda García and Calle Martín, 2012). Finally, we must mention that some natural languages may have other linguistic construction than those used in the English language.…”

Towards a taxonomy of suspected forgery in authorship attribution field

“…New calibration and applications of this index are having particularly fruitful results (see e.g. [20], [54]). …”

“…type-token ratio), the words" length, the repeated segments of words, the position and recursion of specific keywords (see [38], [32], [64], [35], [61] ) for a recent literature review on the topic; see [54] for a presentation and discussion of the different methods as well as the related references. ).…”

“…One of the first tools developed to test in more global terms the lexical similarity between two corpora is the lexical connection index which is the ratio between the shared part of vocabulary and the two corpora joined together, which measures the portion of vocabulary shared by two texts [19,[52][53][54][55].…”

Improving Labbé’s Intertextual Distance: Testing a Revised Version on a Large Corpus of Italian Literature

“…Choosing an appropriate similarity measure is crucial for text clustering, but hundreds of different measures are available (Rudman, 1998), whereas no measure can be considered best suited for all applications. A pairwise measure of similarity should mirror the proximity of two texts but this numeric value depends on the lexical features of texts and on the measure itself (Tuzzi, 2010;Savoy, 2012).…”