Disambiguation Approach for Arabic-English Query Translation in Cross
Language Information Retrieval System
Ahmad M. Hasnah and Jihad M. Jaam
University of Qatar Dept. of Computer Science,
Doha, Qatar, hasna@qu.edu.qa  , hjaam@qu.edu.qa
 

Abstract:
Dictionaries have often been used for query translation in cross-language information retrieval (CUR). However, we are usually faced with translation ambiguity, i.e. multiple translation (meaning) are found in a dictionary/or every word. In this paper we present and evaluate a new method to improve the query translation using a bilingual dictionary. The method is base on user feedback in addition to translation back of sentences from the initially retrieved document using Every-Match approach, to identify the best translation of the user query terms. Our experiments on Arabic-English cross language retrieval show that our technique result in significant improvement over simple dictionary approaches.
Keywords: Query Disambiguation, Cross Language Retrieval, Bilingual Dictionary