Datasets and Models for Authorship Attribution on Italian Personal Writings

TL;DR

This paper introduces new datasets and models for authorship verification on short Italian texts, analyzing how genre, topic, gender, and length influence attribution accuracy, showing AV's feasibility with limited data.

Contribution

It presents novel Italian datasets for authorship verification and analyzes the impact of various factors on attribution performance.

Findings

AV is feasible with limited data

Gender and topic influence attribution results

Controlling for gender and topic is important

Abstract

Existing research on Authorship Attribution (AA) focuses on texts for which a lot of data is available (e.g novels), mainly in English. We approach AA via Authorship Verification on short Italian texts in two novel datasets, and analyze the interaction between genre, topic, gender and length. Results show that AV is feasible even with little data, but more evidence helps. Gender and topic can be indicative clues, and if not controlled for, they might overtake more specific aspects of personal style.