以实现中文本体三元组的快速索引和更新为目的,首先分析中文本体的知识结构,并通过OWL建模语言将其内部的各种知识单元转换成关系三元组和属性三元组构成的知识网络;然后利用《通用规范汉字表》对体现实体间语义信息的关系三元组进行编码,形成主语前缀和宾语前缀索引表,同时对描述实体特征的属性三元组进行编码,形成主语前缀和主语后缀索引表;接下来详细论述关系和属性三元组的更新方法;最后选取多个本体文件对三元组的汉字索引和更新方法进行测试和分析.结果表明,单字索引能够有效地减少本体三元组的更新和检索时间.

This paper focuses on the building and updating methods of Chinese ontology triples index based on Chinese characters. The architecture of Chinese ontology is described firstly, as well as knowledge network consisting of relation triples and property triples through OWL. And then, the building and updating methods about relation and property triples indexing by Chinese characters are introduced in details. The authors make an experiment with a large number of instances in order to test performance of index and update algorithms. The result indicates indexing Chinese triples with Chinese characters can reduce the retrieval and update time. Finally, the article discusses the experimental result and proposes the research emphasis in the future.