OSMAN ― A Novel Arabic Readability Metric

OSMAN – A Novel Arabic Readability Metric Mahmoud El-Haj and Paul Rayson School of Computing and Communications, Lancaster University, UK {m.el-haj, p.rayson}@lancaster.ac.uk Abstract We present OSMAN (Open Source Metric for Measuring Arabic Narratives) - a novel open source Arabic readability metric and tool. It allows researchers to calculate readability for Arabic text with and without diacritics. OSMAN is a modified version of the conventional readability formulas such as Flesch and Fog. In our work we introduce a novel approach towards counting short, long and stress syllables in Arabic which is essential for judging readability of Arabic narratives. We also introduce an additional factor called “Faseeh” which considers aspects of script usually dropped in informal Arabic writing. To evaluate our methods we used Spearman’s correlation metric to compare text readability for 73,000 parallel sentences from English and Arabic UN documents. The Arabic sentences were written with the absence of diacritics and in order to count the number of syllables we added the diacritics in using an open source tool called Mishkal. The results show that OSMAN readability formula correlates well with the English ones making it a useful tool for researchers and educators working with Arabic text. Keywords: Arabic, readability, diacritics, OSMAN, flesch, fog, parallel, corpus, NLP, corpus linguistics 1. Introduction and background Readability measures the ease with which a written text can be understood by readers from different educational levels. Early research by Kitson (1921) showed that sentence and word length were best indicators of a text being easy to read. Other factors include number of characters, the count of hard and complex words as well as the number of syl- lables in each word. Since then several popular readability metrics have been introduced such as Dale-Chall (Dale and Chall, 1948) Fog (Gunning, 1968) and Flesch (Kincaid et al., 1975). Readability formulas predict reading difficulty associated with text. This helps educators in selecting the appropriate text for their audience. The majority of the current read- ability formulas work with text written in English (Gun- ning, 1968; McLaughlin, 1969; Kincaid et al., 1975) by counting the number of syllables in the words. Applying those formulas to other languages is more problematic, es- pecially when syllables are not easy to count. For other languages such as Arabic the only current solution is to use language independent metrics such as Laesbarheds-Index (Lix) (Björnsson, 1968) and Automated Readability Index (ARI) (Smith et al., 1967). These metrics are restricted to other factors such as number of words and sentences in ad- dition to the number of long words1 but do not consider any language specific factors. Arabic is highly inflectional and derivational, therefore word length and number of sen- tences cannot be used as the only indicators of text read- ability (Al-Ajlan et al., 2008). Previous work on Arabic readability and Natural Language Processing has suggested the removal of diacritics to sim- plify the language processing but this results in missing vi- tal information about the syllables leading to pronunciation ambiguity (Habash et al., 2007). …

Content truncated to 500 words