Who wrote 'The World of Saddam Hussein?' A supervised machine learning approach.

This article investigates the authorship of the 2003 Arabic novel *The World of Saddam Hussein* (TWOSH), published under the pseudonym Mahdy Haidar, using supervised machine learning (ML) techniques for authorship attribution (AA). By analyzing a corpus of works from eight suspected authors with two feature sets—top frequent words and top free-standing function word compounds—the study achieves near-perfect classification accuracy, ultimately attributing TWOSH most likely to Lebanese author Rabee Jabir. The research highlights the challenges of AA in Arabic due to its complex morphology and introduces a novel feature extraction method focusing on free function words to avoid morphological segmentation. The findings suggest Jabir's use of a pseudonym may relate to concerns over personal safety and privacy given the novel's controversial portrayal of Saddam Hussein.