A lexical knowledge base approach for English–Chinese cross‐language information retrieval

This study proposes and explores a natural language processing‐ (NLP) based strategy to address out‐of‐dictionary and vocabulary mismatch problems in query translation based English–Chinese Cross‐Language Information Retrieval (EC‐CLIR). The strategy, named the LKB approach, is to construct a lexical knowledge base (LKB) and to use it for query translation. In this article, the author describes the LKB construction process, which customizes available translation resources based on the document collection of the EC‐CLIR system. The evaluation shows that the LKB approach is very promising. It consistently increased the percentage of correct translations and decreased the percentage of missing translations in addition to effectively detecting the vocabulary gap between the document collection and the translation resource of the system. The comparative analysis of the top EC‐CLIR results using the LKB and two other translation resources demonstrates that the LKB approach has produced significant improvement in EC‐CLIR performance compared to performance using the original translation resource without customization. It has also achieved the same level of performance as a sophisticated machine translation system. The study concludes that the LKB approach has the potential to be an empirical model for developing real‐world CLIR systems. Linguistic knowledge and NLP techniques, if appropriately used, can improve the effectiveness of English–Chinese cross‐language information retrieval.