MADAD: A Readability Annotation Tool for Arabic Text

MADAD: A Readability Annotation Tool for Arabic Text Nora Al-Twairesh1, Abeer Al-Dayel2, Hend Al-Khalifa3, Maha Al-Yahya4, Sinaa Alageel5, Nora Abanmy6 and Nouf Al-Shenaifi7 1,2,3,4,7College of Computer and Information Science and 5,6College of Pharmacy King Saud University Riyadh, Saudi Arabia {1twairesh|2aabeer|3hendk|4malyahya|5salageel|6nabanmy|7noalshenaifi}@ksu.edu.sa Abstract This paper introduces MADAD, a general-purpose annotation tool for Arabic text with focus on readability annotation. This tool will help in overcoming the problem of lack of Arabic readability training data by providing an online environment to collect readability assessments on various kinds of corpora. Also the tool supports a broad range of annotation tasks for various linguistic and semantic phenomena by allowing users to create their customized annotation schemes. MADAD is a web-based tool, accessible through any web browser; the main features that distinguish MADAD are its flexibility, portability, customizability and its bilingual interface (Arabic/English). Keywords: Readability, annotation, Arabic NLP 1. Introduction Corpus annotation is defined as "the practice of adding interpretative linguistic information to an electronic corpus" (Garside et al., 1997). It is considered as an added value to the raw corpus and a crucial contribution to it (Garside et al., 1997). Several types of annotations exist in Natural Language Processing (NLP) field, these include: structural annotation, POS tagging, morphological annotation, syntactic annotation, semantic annotation, pragmatic annotation and stylistic annotation. Also new types of annotation are emerging as new research fields are evolving in NLP, such as sentiment annotation and readability annotation. Readability is defined as the degree to which a text can be understood (Klare, 2000). Assessing text readability is a long-established problem which aims to grade the difficulty or the ease of the text. Determining readability level is an important measurement to specify the possible audiences of text materials and to evaluate the impact on the readers. One of the obstacles Arabic readability research faces is the lack of sufficiently large data sets for which annotators provide labels with sufficient readability assessments. The construction of a corpus, which can serve as a gold standard to test new readability prediction tools, is needed. In an effort to address this necessity, we propose MADAD, a collaborative online tool to construct a corpus of readability assessments for the Arabic Language. We named the tool MADAD “ددم” which is an Arabic term that means adding or increasing to an entity, since the main process of this tool is enriching the text with additional context. The readability assessment feature that MADAD offers is flexible in which the readability assessments could be carried out on sentence, phrase and paragraph level. Furthermore, MADAD has two methodologies to construct a corpus of readability assessments: pair-wise comparison and direct evaluation of text difficulty. Up to our knowledge, we are the first to provide the readability annotation feature as a collaborative online tool to help in constructing training corpora for different readability services. On the other hand, MADAD can also be used as a general-purpose annotation tool for Arabic text. Since the tool allows its users to propose their ad-hoc annotation schemes; there are no …

Content truncated to 500 words