Adobe PDF (4.35 MB)
Title Details:
Stylometric Authorship Attribution with more than Two Candidates
Authors: Mikros, George
Reviewer: Markopoulos, Georgios
Subject: HUMANITIES AND ARTS > LINGUISTICS > FIELDS OF LINGUISTICS > DESCRIPTIVE LINGUISTICS > COMPUTATIONAL LINGUISTICS
Description:
Abstract:
In this chapter, we extend the research question of stylometric attribution of authorship to more than two potential authors. Additionally, we examine whether the text is the minimum unit of stylometric measurement or if a subset of it (equal-length sub-segments) can be effectively utilized. The corpus used for the experiments in this chapter belongs to the blog genre and is balanced in terms of the number of posts per author and the author's gender, allowing us to study the effect of the number of authors without the risk of confounding factors. It includes 1,000 posts from 20 authors (10 men and 10 women), with 50 posts from each author. The use of the Support Vector Machines algorithm is introduced, and alternative methodologies for selecting appropriate stylometric features to optimize machine learning for this research problem are presented.
Linguistic Editors: Segkouli, Sofia
Technical Editors: Theodoraki, Alexandra
Graphic Editors: Theodoraki, Alexandra
Type: Chapter
Creation Date: 2015
Item Details:
License: Attribution – NonCommercial – NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Handle http://hdl.handle.net/11419/4865
Bibliographic Reference: Mikros, G. (2015). Stylometric Authorship Attribution with more than Two Candidates [Chapter]. In Mikros, G. 2015. Computational Stylistics [Undergraduate textbook]. Kallipos, Open Academic Editions. https://hdl.handle.net/11419/4865
Language: Greek
Is Part of: Computational Stylistics
Publication Origin: Kallipos, Open Academic Editions