为解决语义网检索过程中缺少推理导致语义检索性能不高的问题,提出一个基于推理的语义网检索模型,并介绍了该模型实现的关键技术。针对构建的领域本体,使用SWRL规则语言进行本体完善把本体中的隐性知识表示出来;在信息抽取过程中,对于识别出的实体,利用推理规则,获得更加丰富的实例知识库;对于用户输入的查询条件,利用完善的本体知识库及规则得到更多的相似概念实现查询扩展;进行语义匹配,获得更为精准的检索结果。实验结果表明,该模型能提高信息检索的语义性,得到较满意的信息检索结果。
To address the problem of lack of reasoning in semantic web retrieval, a semantic web retrieval model based on reasoning is proposed. And some key technologies are introduced. SWRL rule language can perfect the domain ontology and reveal tacit knowledge; reasoning rules can help identified entity to get richer instance knowledge during the information extraction process; ontology and rules can help to realize query expansion. In the end, semantic matching help to get more accurate search results. And the following test shows that the method can get higher performance results.