Weighted Voting and Meta-Learning for Combining Authorship Attribution Methods

Petrovic, Smiljana; Petrovic, Ivan; Palesi, Ileana; Calise, Anthony J.

doi:10.1007/978-3-030-03493-1_35

Cited by 2 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This task can be described in two different configurations: (a) binary classification, where we try to guess whether an author produced a text or not; (b) multi-class classification, where each class corresponds to an author, and we try to attribute a text to one of the authors. Initially, the AA task was approached as a binary classification, and in the case of multiple authors, a voting classifier was implemented [23,24]. In this configuration, the basic models were limited to two authors, and the more fluent ones had to support a voting mechanism with multiple authors.…”

Section: Background and Related Workmentioning

confidence: 99%

Shedding Light on the Dark Web: Authorship Attribution in Radical Forums

Ranaldi

Fallucchi

et al. 2022

Information

View full text Add to dashboard Cite

Online users tend to hide their real identities by adopting different names on the Internet. On Facebook or LinkedIn, for example, people usually appear with their real names. On other standard websites, such as forums, people often use nicknames to protect their real identities. Aliases are used when users are trying to protect their anonymity. This can be a challenge to law enforcement trying to identify users who often change nicknames. In unmonitored contexts, such as the dark web, users expect strong identity protection. Thus, without censorship, these users may create parallel social networks where they can engage in potentially malicious activities that could pose security threats. In this paper, we propose a solution to the need to recognize people who anonymize themselves behind nicknames—the authorship attribution (AA) task—in the challenging context of the dark web: specifically, an English-language Islamic forum dedicated to discussions of issues related to the Islamic world and Islam, in which members of radical Islamic groups are present. We provide extensive analysis by testing models based on transformers, styles, and syntactic features. Downstream of the experiments, we show how models that analyze syntax and style perform better than pre-trained universal language models.

show abstract

Section: Background and Related Workmentioning

confidence: 99%

Shedding Light on the Dark Web: Authorship Attribution in Radical Forums

Ranaldi

Fallucchi

et al. 2022

Information

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

“…In recent years, it has been demonstrated that an ensemble of classifiers, using a broad selection of stylistic features, significantly outperforms individual classifier/feature pairs (Petrovic I. 2019, Petrovic S. 2018). This has motivated research into new, non-traditional stylistic features, which complement the primary features listed above and strengthen the overall accuracy of the attribution hypothesis.…”

Section: Introductionmentioning

confidence: 99%

A Multi-Dictionary Approach to Abstractness/Concreteness-Based Authorship Attribution

Ivanov

2023

FLAIRS

View full text Add to dashboard Cite

We present some early results from a research project aimed at exploring the usefulness of abstractness/concreteness as stylistic features for authorship attribution. We conjecture that authors use abstract/concrete words and phrases in suf- ficiently unique ways, so that machine learning classifiers can learn to distinguish the individual authors’ writing styles. Our approach is based on using the abstractness rat- ings of words and phrases from texts with established au- thorship to generate training vectors for different machine learning classifiers. The combined word/phrase ratings are extracted from two separate abstractness dictionaries – an approach that yields stronger results than using single ab- stractness dictionaries. The paper describes the details of our methodology and compares the results to those obtained using traditional authorship attribution stylistic features. The limitations of our current methodology and directions for further research are outlined at the end of the paper.

show abstract

Weighted Voting and Meta-Learning for Combining Authorship Attribution Methods

Cited by 2 publications

References 11 publications

Shedding Light on the Dark Web: Authorship Attribution in Radical Forums

Shedding Light on the Dark Web: Authorship Attribution in Radical Forums

A Multi-Dictionary Approach to Abstractness/Concreteness-Based Authorship Attribution

Contact Info

Product

Resources

About