arXiv:2604.19499v4 Announce Type: replace 
Abstract: This article introduces two new measures for authorship attribution - Rank-Turbulence Delta and Jensen-Shannon Delta - which generalise Burrows's classical Delta by applying distance functions designed for probabilistic distributions. We first set out the theoretical basis of the measures, contrasting centred and uncentred z-scoring of word-frequency vectors and re-casting the uncentred vectors as probability distributions. Building on this representation, we develop a token-level decomposition that renders every Delta distance numerically interpretable, thereby facilitating close reading and the validation of results. The effectiveness of the methods is assessed on four literary corpora in English, German, French and Russian. The English, German and French datasets are compiled from Project Gutenberg, whereas the Russian benchmark is the SOCIOLIT corpus containing 639 works by 89 authors spanning the eighteenth to the twenty-first centuries. Rank-Turbulence Delta attains attribution accuracy comparable with Cosine Delta; Jensen-Shannon Delta consistently matches or exceeds the performance of canonical Burrows's Delta. Finally, several established attribution algorithms are re-evaluated on the extended SOCIOLIT corpus, providing a realistic estimate of their robustness under pronounced temporal and stylistic variation.

يقدم المقال مقاييس جديدة لنسب المؤلفين، وهي Rank-Turbulence Delta و Jensen-Shannon Delta، والتي تعزز من Delta الكلاسيكي لبوروز من خلال استخدام دوال المسافة المصممة للتوزيعات الاحتمالية. تستند هذه المقاييس إلى إطار نظري يقارن بين طرق التقييم المختلفة ويعيد تفسير متجهات تكرار الكلمات كالتوزيعات الاحتمالية.

El artículo presenta dos nuevas medidas para la atribución de autoría, Rank-Turbulence Delta y Jensen-Shannon Delta, que mejoran el Delta clásico de Burrows al utilizar funciones de distancia diseñadas para distribuciones probabilísticas. Estas medidas se basan en un marco teórico que contrasta diferentes métodos de puntuación z y reinterpreta los vectores de frecuencia de palabras como distribuciones de probabilidad.

L'article présente deux nouvelles mesures pour l'attribution d'auteur, le Rank-Turbulence Delta et le Jensen-Shannon Delta, qui améliorent le Delta classique de Burrows en utilisant des fonctions de distance adaptées aux distributions probabilistes. Ces mesures reposent sur un cadre théorique qui contraste différentes méthodes de z-scoring et réinterprète les vecteurs de fréquence de mots comme des distributions de probabilité.

The article introduces two new measures for authorship attribution, Rank-Turbulence Delta and Jensen-Shannon Delta, which enhance Burrows's classical Delta by utilizing distance functions tailored for probabilistic distributions. These measures are grounded in a theoretical framework that contrasts different z-scoring methods and reinterprets word-frequency vectors as probability distributions.

Rank-Turbulence Delta and Interpretable Approaches to Stylometric Delta Metrics

Was this article worth reading? Share it

Ready to build your own newsroom?