基于领域本体的生物医学语义检索机制研究——以GoPubMed和SEGoPubMed为例

Research of Biomedical Semantic Retrieval Based On Domain Ontology ——A Case Study of GoPubMed and SEGoPubMed

基于领域本体对大型文本语料库的语义检索研究成为近年来信息检索领域研究的热点。GoPubMed和SEGoPubMed作为该领域研究成果的杰出代表，为PubMed存在的问题提供了良好的解决方案。对GoPubMed和SEGoPubMed语义检索机制进行深入研究后显示，GoPubMed应用基于语义网的语义分类工具——_GO和MeSH，将PubMed检索结果统计并对应到相应的分类导航目录，使用户利用该导航便可以快速找到自己所需要的文献，但其仍是基于布尔逻辑模型的字符串匹配机制；而比GoPubMed性能更优越的SEGoPubMed应用潜在语义分析（LSA）技术开创了语义敏感地浏览、连接大型文本语料库与本体的新模式，将具有更广泛的应用前景。

Domain ontology based semantic retrieval of the large-scale text corpus has become a hot area of information retrieval research in recent years. GoPubMed and SEGoPubMed, as an outstanding representative of this research area, provide good solutions to the existing problems of PubMed. Our deep studies on semantic search mechanism of GoPubMed and SEGoPubMed have shown that GoPubMed applies Semantic Web-based classification instruments---GO and MESH, measures and corresponds the PubMed search results to the classification of the corresponding navigation directory so that users can use the navigation to quickly find the literature they need, however, it is still based on the string matching mechanism of Boolean logic model. SEGoPubMed,which performaces better than GoPubMed, applies Latent Semantic Analysis （LSA） technology to create a new paradigm to semantic-sensitive ontology based browsing and linking of large corpus to ontologies. The analysis suggests that the SEGoPubmed performs better than the classic GoPubmed as it incorporates semantics. SEGol？ubMed has a brighter future.