A task oriented extension of the Chinese MeSH Concepts Hierarchy

The Medical Subject Headings (MeSH) thesaurus is used not only as a vocabulary for indexing and cataloguing biomedical articles, but also constitutes an important linguistic resource for text mining in the biomedical domain. As part of our research, the Chinese translation of English MeSH, Chinese Medical Subject Headings (CMeSH), will be used to provide essential terms and concepts to a Chinese-English cross-lingual information retrieval system. The original MeSH uses a small, accurate and concise set of terms designed for indexing and cataloguing. Such a set of terms does not, however, lend itself well to information retrieval applications, in which the ability to expand queries using term synonyms is important to maximise relevant search results. In this paper, we propose a new approach to extending the MeSH concept hierarchy (the MeSH Tree) based on an online version of the CMeSH term list. We use Google to collect synonyms for Chinese terms and calculate weights for each Chinese term and its synonyms. Our extension has been evaluated on a Chinese-English biomedical information retrieval application. The results indicate that the extension of CMeSH improves the performance of the information retrieval application. Furthermore, the extended resource should also be helpful in other related research.