An algorithm for automated authorship attribution using neural networks

Tearle, Matt; Taylor, Kye; Demuth, Howard B.

doi:10.1093/llc/fqn022

Cited by 12 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An example of a stylistic marker is the use of 'txting' shorthand such as 2nite 4 and the use of emoticons such as :-) [6]. It has been shown repeatedly in the literature that determining the particular author from a set of candidate authors is possible by looking at the documents that each author has written and matching a new document of unknown origin to a profile built of each author [14], [23], [26], [29], [32], [35]. This process is known as authorship attribution and is part of the field of authorship analysis which includes author profiling [4], similarity detection [22] and authorship intent determination [17].…”

Section: Introductionmentioning

confidence: 99%

Authorship Attribution for Twitter in 140 Characters or Less

Layton

Watters

Dazeley

2010

2010 Second Cybercrime and Trustworthy Computing Workshop

View full text Add to dashboard Cite

Authorship attribution is a growing field, moving from beginnings in linguistics to recent advances in text mining. Through this change came an increase in the capability of authorship attribution methods both in their accuracy and the ability to consider more difficult problems. Research into authorship attribution in the 19 ℎ century considered it difficult to determine the authorship of a document of fewer than 1000 words. By the 1990s this values had decreased to less than 500 words and in the early 21 century it was considered possible to determine the authorship of a document in 250 words. The need for this ever decreasing limit is exemplified by the trend towards many shorter communications rather than fewer longer communications, such as the move from traditional multi-page handwritten letters to shorter, more focused emails. This trend has also been shown in online crime, where many attacks such as phishing or bullying are performed using very concise language. Cybercrime messages have long been hosted on Internet Relay Chats (IRCs) which have allowed members to hide behind screen names and connect anonymously. More recently, Twitter and other short message based web services have been used as a hosting ground for online crimes. This paper presents some evaluations of current techniques and identifies some new preprocessing methods that can be used to enable authorship to be determined at rates significantly better than chance for documents of 140 characters or less, a format popularised by the micro-blogging website Twitter 1 . We show that the SCAP methodology performs extremely well on twitter messages and even with restrictions on the types of information allowed, such as the recipient of directed messages, still perform significantly higher than chance. Further to this, we show that 120 tweets per user is an important threshold, at which point adding more tweets per user gives a small but non-significant increase in accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

Authorship Attribution for Twitter in 140 Characters or Less

Layton

Watters

Dazeley

2010

2010 Second Cybercrime and Trustworthy Computing Workshop

View full text Add to dashboard Cite

show abstract

“…This is somewhat akin to exploiting a committee (or ensemble) of classifiers, an approach which has sometimes proven more effective than single classifiers (Adamovic et al, 2019;Potha and Stamatatos, 2019). In the setting of AA, Tearle et al (2008) have employed a similar committee of NNs, each one with a different feature set; however, instead of hard-counting the vote of each NN, as they do, we choose a "softer" approach and average the probabilities computed by each branch, akin to what is done by Muttenthaler et al (2019). A scheme of our NN architecture is displayed in Figure 1.…”

Section: Neural Networkmentioning

confidence: 99%

Syllabic Quantity Patterns as Rhythmic Features for Latin Authorship Attribution

Corbara¹,

Moreo²,

Sebastiani³

2021

Preprint

View full text Add to dashboard Cite

It is well known that, within the Latin production of written text, peculiar metric schemes were followed not only in poetic compositions, but also in many prose works. Such metric patterns were based on so-called syllabic quantity, i.e., on the length of the involved syllables, and there is substantial evidence suggesting that certain authors had a preference for certain metric patterns over others. In this research we investigate the possibility to employ syllabic quantity as a base for deriving rhythmic features for the task of computational authorship attribution of Latin prose texts. We test the impact of these features on the authorship attribution task when combined with other topic-agnostic features. Our experiments, carried out on three different datasets, using two different machine learning methods, show that rhythmic features based on syllabic quantity are beneficial in discriminating among Latin prose authors.

show abstract

“…In the last decade, the interest in the Elizabethan playwrights has not faded. Recent work on Marlowe and Shakespeare by Tearle et al ( 2008 ) highlight that Shakespeare was a collaborator on Titus Andronicus, but that it was easy to separate Shakespeare from Marlowe using neural networks. Craig and Kinney ( 2009 ) suggest that there is doubt about the authorship of Henry VI and that Parts 1 and 2 are Marlowe's and not Shakespeare's.…”

Section: Introductionmentioning

confidence: 99%

Using Shakespeare's Sotto Voce to Determine True Identity From Text

2018

View full text Add to dashboard Cite

Little is known of the private life of William Shakespeare, but he is famous for his collection of plays and poems, even though many of the works attributed to him were published anonymously. Determining the identity of Shakespeare has fascinated scholars for 400 years, and four significant figures in English literary history have been suggested as likely alternatives to Shakespeare for some disputed works: Bacon, de Vere, Stanley, and Marlowe. A myriad of computational and statistical tools and techniques have been used to determine the true authorship of his works. Many of these techniques rely on basic statistical correlations, word counts, collocated word groups, or keyword density, but no one method has been decided on. We suggest that an alternative technique that uses word semantics to draw on personality can provide an accurate profile of a person. To test this claim, we analyse the works of Shakespeare, Christopher Marlowe, and Elizabeth Cary. We use Word Accumulation Curves, Hierarchical Clustering overlays, Principal Component Analysis, and Linear Discriminant Analysis techniques in combination with RPAS, a multi-faceted text analysis approach that draws on a writer's personality, or self to identify subtle characteristics within a person's writing style. Here we find that RPAS can separate the known authored works of Shakespeare from Marlowe and Cary. Further, it separates their contested works, works suspected of being written by others. While few authorship identification techniques identify self from the way a person writes, we demonstrate that these stylistic characteristics are as applicable 400 years ago as they are today and have the potential to be used within cyberspace for law enforcement purposes.

show abstract

An algorithm for automated authorship attribution using neural networks

Cited by 12 publications

References 16 publications

Authorship Attribution for Twitter in 140 Characters or Less

Authorship Attribution for Twitter in 140 Characters or Less

Syllabic Quantity Patterns as Rhythmic Features for Latin Authorship Attribution

Using Shakespeare's Sotto Voce to Determine True Identity From Text

Contact Info

Product

Resources

About