Automatic Wordnet Development for Low-Resource Languages using Cross-Lingual WSD

Wordnet is an effective resource in natural language processing and information retrieval, especially for semantic processing and meaning related tasks. So far wordnet has been constructed in many languages. However, automatic development of wordnet for lowresource languages has not been studied well. In this paper an Expectation-Maximization algorithm is used to train high quality and large scale wordnet for resource-poor languages. The proposed method benefits from cross-lingual word sense disambiguation and develops a wordnet just using a bilingual dictionary and a monolingual corpus. The proposed method has been executed on Persian as a resource-poor language and the resulting wordnet has been evaluated through several experiments. Results show that the induced wordnet has a precision of 90% and recall of 35%.