基于语义的文档标注方法研究  

Research on semantic-based annotation for documents

作　　者：张泽宇 李莉[1] 谭凤[1] 晏萍丽[1] 

机构地区：[1]西南大学计算机与信息科学学院,重庆400715

出　　处：《计算机工程与科学》2013年第9期151-156,共6页Computer Engineering & Science

基　　金：国家自然科学基金资助项目(61170192)

摘　　要：针对现有语义标注系统效率低下的现状,借鉴NCBO Annotator的思想,提出了一种基于语义的文档标注方法,旨在结合本体知识库和WordNet的语义知识,构建具有丰富语义描述能力的字典库,并利用字典库实现文档的语义标注。实验表明,本方法能够提高搜索资源的语义,进而提高搜索的准确度。此外,利用本方法还可提供语义扩展搜索。最后将字典库和标注过程以RESTful Web Service的方式呈现,极大地方便了广大用户对字典库和语义资源的使用。Abstract.To address the issue that existing semantic-based annotation system has low efficiency, a semantic-based annotation approach is proposed based on the idea of NCBO annotator. The purpose is to create the （semantic） dictionary with strong semantic description ability by combining Ontology Knowl- edge Base and WordNet Semantic Base, then use the dictionary to support semantic-based annotation. The experiment based on the data obtained from popular domains demonstrates the effectiveness and proficiency of the proposed approach in terms of high precision and recall measure. The approach, im- plemented as RESTful Web service, provides not only a greater level of convenience for further reference but also performance-enhancing potential across application domains.

关 键 词：本体 WORDNET 字典库 语义标注 RESTFUL Web SERVICE 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]