An Arabic WordNet Enrichment Approach using Machine Translation and External Linguistic Resources

The development of Arabic WordNet (AWN) has provided the Arabic Natural Language Processing (NLP) community with a lexical resource. However, due to the considerable lack of AWN Synsets compared to other WordNets, the use of this resource in Arabic NLP applications cannot be considered as an option yet. In this paper, we propose an automatic approach for AWN enrichment, based on three features: (1) the translation of Princeton WordNet (PWN) words and their direct Hypernyms; (2) an automatic validation of the hypernymy relation that is suggested as existing between words using Wikipedia articles; and (3) the integration of validated Synsets in AWN including their definitions, examples and synonyms that get extracted using dictionaries and other resources. The proposed approach has been implemented and a preliminary evaluation shows that the obtained results are promising.