Rule-based Annotation Tools for Modern Standard Arabic

Rule-based Annotation Tools for Modern Standard Arabic, an IRCSET EMPOWER Initiative-funded postdoctoral fellowship, is a 2 year project (December 2009 - December 2011) with 1 academic partner (Dublin City University).

In this research Dr. Mohammed Attia proposes the development of a suite of annotation tools for unrestricted Modern Standard Arabic (MSA) text using Finite State Morphology (FSM) and Constraint Grammar (CG) formalisms including morphological analysis, lemmatization, tokenization, part-of-speech (POS) tagging and partial parsing. In order to develop these tools, a representative corpus of MSA texts will be created, and a gold standard will be manually annotated for development and evaluation purposes. His work can support deep parsing of free text, as for example with the ATB-based LFG grammars which need a lot of morphological information to generate the required f-structure annotations. Deep parsing is required for meaning sensitive applications that analyse search queries, index documents and general semantic representations.

Please contact

• Dr. Mohammed Attia