Amr&MohamedSabaa at AraGenEval shared task: Arabic Authorship Identification using Term Frequency – Inverse Document Frequency Features with Supervised Machine Learning
Year: 2025
Type: conference-paper
Abstract: This paper presents our approach to the Ara-GenEval 2025 shared task on Arabic authorship attribution (Task 2).We developed an enhanced traditional machine learning system that combines word-level and character-level TF-IDF features with multiple classification algorithms.Our system achieved 88.90% accuracy and 82.74% macro F1-score on the official test set using Logistic Regression.During development, we evaluated multiple models on the validation set, where Linear SVM achieved the highest performance with 93.22% accuracy and 87.52% macro F1-score.The approach demonstrates the effectiveness of feature engineering and proper text preprocessing for Arabic authorship attribution tasks without relying on deep learning architectures.
Authors: Amr Sabaa, Mohamed Sabaa
Institutions: Cairo University, Najran University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4416037832