Moroccan Arabic Vocabulary Generation Using a Rule-Based Approach (6. Conclusion) - tachicart

6. Conclusion


In this paper, we have presented MORV the first NLP resource that targets Moroccan morphology. MORV is a flexible, FST-based tool for automatically generating Moroccan Arabic dialectal words. The FST topology is inspired by earlier approaches used to generate Arabic words. Additionally, we have exploited Moroccan Arabic resources for this purpose and generated 4.68M Moroccan words with full morphological attributes. We have also shown how important keeping the linguistic knowledge separated from the processing algorithm in order to ensure MORV extensibility. MORV evaluation shows that it reaches a precision of 94% and covers 84% of the Moroccan Arabic used in social media. These rates can be improved by incorporating new lexicon entries and running the corresponding generation process which is the subject of our future work.

Next section