Key points are not available for this paper at this time.
We present a diacritization system for written Arabic which is based on a lexical resource. It combines a tagger and a lexeme language model. It improves on the best results reported in the literature.
Habash et al. (Mon,) studied this question.