Middle East Studies Lecture
Machine Learning: Linguistic Cues for Arabic Authorship Analysis
- Date
- Tuesday 7 April 2026
- Time
- Serie
- Middle East Studies Lectures
- Address
-
Herta Mohr
Witte Singel 27A
2311 BG Leiden - Room
- 1.128 (Verbarium)
As early as the 1960s, numerical approaches to authorship analysis have been used to investigate historical documents. A famous example is the use of statistical methods of textual properties such as common word frequencies to provide evidence to the attribution of question essays in the Federalist Papers. In this [talk/presentation?], Hossam Ahmed explores extreme cases where the question is to verify the authorship of a document to a single author, or where the number of authors of a set of documents is unknown. The rich morphology and flexible word order allow Machine Learning algorithms a large array of linguistic features that can be used for authorship analysis, and reveal what makes each author unique.