15/05/2026
مقالات مجلد 13 عدد 1 أبريل 2026
--------------------------------------
Hybrid Machine Learning Framework for Authorship Identification Based on Writing Style
نوع المستند : المقالة الأصلية
المؤلفون
heba marie 1 Mohamed abd alhady 2 hendy ahmed 3
1 كلية الآداب جامعة حلوان
2 كلية الآداب جامعة القاهرة
3 Damietta university
10.21608/jesi.2025.433517.1160
المستخلص
Authorship identification software is a computational tool that aims to identify the author of a given text. It is based on the idea that everyone has a unique writing. This study presents the development of an intelligent authorship identification system that leverages advanced machine learning and deep learning techniques to determine the authorship of given text documents. The system incorporates three primary models—Long Short-Term Memory (LSTM), Bidirectional Encoder Representations from Transformers (BERT), and Support Vector Machines (SVM)—each designed to analyze distinctive linguistic and stylistic features of individual authors. The expected classification accuracies for the models were set as follows: 70% for LSTM, 75% for BERT, and 90% for SVM. These benchmarks guided the design, optimization, and evaluation of each algorithm.
The research methodology followed the Agile software development model, allowing for iterative refinement and continuous integration of user feedback throughout the system's implementation. The system architecture includes modular components for preprocessing, model inference, and a user-friendly interface for document upload and authorship prediction. Experiments were conducted using benchmark datasets of labeled textual data, and each model was assessed based on prediction accuracy and overall system performance.
Although actual results slightly trailed expectations, the system achieved strong performance: 63% with LSTM, 71% with BERT, and 86% with SVM, validating the effectiveness of the proposed approach. This tool has broad applications in forensic linguistics, academic integrity enforcement, and digital content verification. The study contributes to the advancement of authorship attribution techniques and offers a scalable, accurate,
الكلمات الرئيسية
Authorship Identification Writing Style Analysis Plagiarism Detection Literary Attribution
