Authorship Attribution in Arabic Poetry
No Thumbnail Available
Date
2018
Authors
Balatiah, Razan
Mahmoud, Sara
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This project represents an authorship attribution in Arabic poetry using Ara- bic Natural Language Processing (ANLP) and machine learning techniques. The main objective of this project is to employ computer software in the treatment of Arabic poetry so that the machine , according to a given poem or a part of it with an unknown poet , can automatically predict the poet who wrote the poem . Our project targets the young poets , students of Arabic literature and poetry lovers to know the styles of poets ,to acquire the wisdom ,the messages and the morals , and also to spread Arabic poetry which has deep importance and finally to avoid Literary thefts. The main challenge is: Can we identify the original author for unknown text among a set of candidate authors automatically by the machine? To do that we will use style markers and features to identify the author such as the charac- ters, length of sentences and words, meter, rhyme and dicritization (harakat) in the poems. All these features are used as input data for classification algorithms . Besides, dataset of poems with known poets will be used as training data , and the data of texts whose authors are unknown (text in the test dataset) will be mined to find out the the writer’s style which indicates his/her name. This project has been done on a group of 8 authors and the results with classification precision was 83%