Developing easy-to-read written medicine information continues to be a challenge in health communication. Readability aims to gauge the difficulty level of a text. Various formulas and machine learning algorithms have proposed to judge the readability of health materials and assist writers in identifying possible problems related to text difficulty. For this reason, having corpus annotated with readability levels is fundamental to evaluating the readability formulas and training machine learning algorithms. Arabic suffers from a lack of annotated corpora to evaluate text readability, especially for health materials. To address this shortage, we describe a baseline results towards constructing readability Corpus ARC-WMI, a new Arabic collection of written medicine information annotated with readability levels. We compiled a corpus of 4476 sentences with over 61k words, extracted from 94 sources of Arabic written medicine information. These sentences were manually annotated and assigned a readability level (“Easy,” “Intermediate,” or “Difficult”) by a panel of five health-care professionals.
@InProceedings{AL-DAYEL18.9, author = {Abeer AL-Dayel ,Hend Al-Khalifa ,Sinaa Alaqeel ,Norah Abanmy ,Maha Al-Yahya and Mona Diab}, title = {ARC-WMI: Towards Building Arabic Readability Corpus for Written Medicine Information}, booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {may}, date = {7-12}, location = {Miyazaki, Japan}, editor = {Hend Al-Khalifa and King Saud University and KSA
Walid Magdy and University of Edinburgh and UK
Kareem Darwish and Qatar Computing Research Institute and Qatar
Tamer Elsayed and Qatar University and Qatar}, publisher = {European Language Resources Association (ELRA)}, address = {Paris, France}, isbn = {979-10-95546-25-2}, language = {english} }
Summary of the paper