Using Ontology for Revealing Authorship Attribution of Arabic Text
Abeer H. El Bakly1, Nagy Ramadan Darwish2, Hesham A.Hefny3
1Abeer H. El Bakly *, Information system and technology department, Faculty of Graduate Studies for Statistical Research, Cairo University.
2Nagy Ramadan Darwish, Information system and technology department, Faculty of Graduate Studies for Statistical Research, Cairo University.
3Hesham A.Hefny, computer system department, Faculty of Graduate Studies for Statistical Research, Cairo University,
Manuscript received on April 05, 2020. | Revised Manuscript received on April 25, 2020. | Manuscript published on April 30, 2020. | PP: 143-151 | Volume-9 Issue-4, April 2020. | Retrieval Number: C6412029320/2020©BEIESP | DOI: 10.35940/ijeat.C6412.049420
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Authorship attribution analysis is a research field that assigns an author to an unknown text based on writing features. These features reflect the author’s gender, age, religion, education, job, motivation or ideology. It has several types of features such as character, lexical, Syntactic, Structural and Semantic. This research proposed using Arabic ontology as a semantic feature in authorship attribution through a proposed new model. In the Islamic society, there is a problem in detecting unknown fatwa to specific jurisprudence doctrine so this research proposed a new model for detecting unknown fatwa to specific jurisprudence doctrine. This model depends on a new corpus which is manually collected and annotated fatwas from books of Islamic jurisprudence doctrines. This corpus is called ElWafaa LlFokahaa. It includes the fatwas of traveller’s prayer for main Islamic doctrines (Hanfi, Shafie, Malki, and Hanbali). The proposed model used Arabic ontology for traveller’s prayer in each Islamic doctrine which is established with protégé framework. It is divided into a training set in 70% of fatwas (known fatwas the owing Islamic jurisprudence doctrines) and 30% testing set (unknown fatwas of Islamic jurisprudence doctrines). For evaluating the proposed model, it is used the proposed evaluated method which is 90% with final experiments.
Keywords: Artificial intelligence, fiqh, similarity, feature selection, ontology, authorship attribution.