文摘
The rapid growth of information sources has produced a large amount of electronically stored documents evolving every day. The development of Information Retrieval Systems (IRS) is a response to this growth, which aims to help the user identify relevant information. Recent IRS proposes to guide the user through providing domain knowledge in the form of controlled vocabularies or terminologies and thus, domain ontologies. In this context, it is necessary to develop multilingual termino-ontological resources. This paper proposes a new approach for bilingual domain terminology extraction in Arabic and English languages as a first step in the bilingual domain ontology building, to be exploited in terminological search. The approach uses arabic vocalized texts to reduce ambiguities and the alignment process to extract the english translations. To the best of our knowledge, the process implemented in our approach (morphological analysis, arabic terminology extraction, alignment and extraction of english translations) is the first work in the field of arabic-english bilingual domain terminology extraction. The results of experiments are encouraging showing rates of relevant term extraction from multiple domains and their translations exceeding 89%.