Abstract
We present CALIMAGLF, a Gulf Arabic morphological analyzer currently covering over 2,600 verbal lemmas. We describe in detail the process of building the analyzer starting from phonetic dictionary entries to fully inflected orthographic paradigms and associated lexicon and orthographic variants. We evaluate the coverage of CALIMAGLF against Modern Standard Arabic and Egyptian Arabic analyzers on part of a Gulf Arabic novel. CALIMAGLF verb analysis token recall for identifying correct POS tag outperforms both the Modern Standard Arabic and Egyptian Arabic analyzers by over 27.4% and 16.9% absolute, respectively.
Cite
CITATION STYLE
Khalifa, S., Hassan, S., & Habash, N. (2017). A Morphological Analyzer for Gulf Arabic Verbs. In WANLP 2017, co-located with EACL 2017 - 3rd Arabic Natural Language Processing Workshop, Proceedings of the Workshop (pp. 35–45). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/w17-1305
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
