Résumé:
Authorship verification (AV) is one of various topics parts of authorship analysis field that deals
with the problem of determine whether two texts were written by the same author or not. A
combination of a similarity-based methods and relevant linguistic features are used to achieve
high accuracy authorship verification. To address this problem, we proposed a new approach
using the Autoencoder deep learning method. Challenges in the context of Authorship
verification have greatly increased in recent years, as the challenge of PAN (series of scientific
events) for three last years from 2020 to 2022. To experiment our approach, we used the data
provided by PAN 2021 AV task.
Keywords: Authorship verification, PAN, Autoencoder, deep learning, similarity.
