Improved Written Arabic Word Parsing through Orthographic, Syntactic and Semantic constraints

The script-based and morphological characteristics of the Arabic language increase considerably the number of alternative analyses output by any morphological parser that does not use orthographic, syntactic and semantic constraints. In order to reduce time-wasting and error-prone proliferation of multiple outputs to be filtered in a post-processing phase, we have tried to optimize word processing by providing the morphological parser with multiple levels of information. We have operated at three such levels: orthography, morpho-syntax and semantics.