Cross Language (Thai-English) Information Retrieval: Concepts and Challenges

Abstract

Documents on the Internet have been written using several languages. The benefit of this is those documents are useful for users to verify the information from different sources. However, users are not able to use a single language in a query to retrieve all relevant documents written in different languages. Moreover, some users do not know exactly what the keywords to be used in a query to retrieve desired documents. As a result, search engine cannot find the relevant documents effectively. In addition, a keyword can refer to many different concepts in the real world, so-called “Polysemy” or many keywords refer to one thing, socalled “Synonym”. They are two significant problems that decrease a search engine performance. Consequently, many researchers try to overcome the problem by developing the Cross-Language Information Retrieval (CLIR) system in order to retrieve documents written by different languages from using a single query. This idea is now a new trend of search engine and can be developed as a commercial product for a popular search engine e.g. Google or Bing. The article presents the concepts and ideas of CLIR and summary of the main challenges in this research area.