Towards the development of an automatic readability measurements for arabic language

Abstract

Currently, there are more than 200 readability formulas developed since the 1920s. Only a handful of these formulas are reliable to determine the reading-level of a sample text. Ascertaining the readability of curricula is an important step toward optimizing the effectiveness of the educational progress. Readability measurements are done using manual computation. This is a tedious and time-consuming task. However, nowadays, automatic readability computation has gained much popularity than ever before. This can be attributed to the advancement in the field of natural language processing. Languages such as English and Spanish have benefited from the formation of automatic readability systems. Absence of automated readability measurement of Arabic texts and the large amount of information which are written in Arabic encouraged us to work on finding an automatic readability system for our language; the mother language for millions of Arabs. In this paper we will review the different readability research and then propose a system for automating the readability measurement of Arabic text. Within the paper we also report the results of our pilot experiment carried out on the different well-known Arabic, Swedish and English readability formulas.