Automatic extraction of lemma-based bilingual dictionaries for morphologically rich languages
Saleh, Ibrahim Mohamed Hassan.
Thesis (M.S.)--Georgetown University, 2009.; Includes bibliographical references.; Text (Electronic thesis) in PDF format. The present study introduces a system designed to automatically extract a comprehensive up-to-date Arabic-English lemma-based MRD from parallel corpora. The designed system makes use of state-of-the-art techniques in both aligning parallel corpora and deciding which alignment pairs are highly probable good MRD entries. Comparing the results of the present system to Buckwalter's manually built dictionary shows that the automatic extraction of lemma-based bilingual MRD is better in terms of time, coverage, dialectal issues, and updating.