Towards Web Mining of Query Translations for Cross-Language Information Retrieval in Digital Libraries

This paper proposes an efficient client-server-based query translation approach to allowing more feasible implementation of cross-language information retrieval (CLIR) services in digital library (DL) systems. A centralized query translation server is constructed to process the translation requests of cross-lingual queries from connected DL systems. To extract translations not covered by standard dictionaries, the server is developed based on a novel integration of dictionary resources and Web mining methods, including anchor-text and search-result methods, which exploit huge amounts of multilingual and wide-scoped Web resources as live bilingual corpora to alleviate translation difficulties, and have been proven particularly effective for extracting multilingual translation equivalents of query terms containing proper names or new terminologies. The proposed approach was implemented in a query translation engine called LiveTrans, which has been shown its feasibility in providing efficient English-Chinese CLIR services for DL.