基于本体的社保审计知识库的构建研究

发布时间：2018-06-06 01:52

本文选题：本体构建 + 本体扩充　；参考：《哈尔滨工程大学》2011年硕士论文

【摘要】：本体在知识表示中的应用越来越广泛,很多专家、学者对本体构建技术进行了深入的研究,相继构建了一些通用本体库,但通用本体库中的专业领域术语却比较匮乏,为了满足特定领域的需求,就需要构建面向领域的本体库。本文研究的主要内容就是构建社会养老保险领域的本体库。为了解决社保领域文献众多、组织管理复杂的问题,本文利用本体作为知识表达和存储的方式。构建本体时,如果完全使用手工的方式会使本体构建的过程比较复杂,如果完全使用统计的方式会影响本体的构建效果,本文结合了这两种方法来完成本体的构建,首先通过人工方式构建本体的框架,保证了本体结构的相对准确性,其次使用本体的属性集或操作对象集的包含关系实现本体的自动扩充。在进行本体自动扩充的过程中,本文建立了本体的属性集和操作对象集,然后分八种情况讨论了经过文本分词、关键词抽取、片段抽取得到的主、谓、宾三元组片段插入本体的方法,通过对这八种情况进行分析,提炼出操作对象集的更新方法,随后,使用操作对象集的包含关系在本体中确定新节点的插入范围,并在插入范围内使用概念相似度计算公式确定具体的插入位置,最终把领域本体从领域文献中提炼出来。实验结果表明这种本体自动构建的过程可以有效的帮助领域专家进行社保本体的构建,并对其它领域本体的构建也有相当的参考价值。
[Abstract]:Ontology is more and more widely used in knowledge representation. Many experts and scholars have carried out in-depth research on ontology construction technology, and constructed some general ontology libraries one after another, but the terminology in general ontology database is relatively scarce. In order to meet the needs of a specific domain, it is necessary to construct a domain-oriented ontology library. The main content of this paper is to construct the ontology library in the field of social endowment insurance. In order to solve the problem of numerous documents and complicated organization management in social security field, this paper uses ontology as a way to express and store knowledge. When building ontology, the process of ontology construction will be complicated if the manual way is used completely, and the effect of ontology construction will be affected if the method of statistics is used completely. In this paper, we combine these two methods to complete the construction of ontology. Firstly, the framework of ontology is constructed manually to ensure the relative accuracy of ontology structure. Secondly, the automatic extension of ontology is realized by using the property set of ontology or the inclusion relation of operation object set. In the process of automatically extending ontology, this paper establishes the attribute set and operation object set of ontology, and then discusses the main object, which is obtained by text partitioning, keyword extraction and fragment extraction, in eight cases. By analyzing the eight cases, the updating method of the operation object set is extracted. Then, the inclusion relation of the operation object set is used to determine the insertion range of the new node in the ontology. Finally, the domain ontology is extracted from the domain literature by using the concept similarity calculation formula to determine the specific insertion position. The experimental results show that the process of automatic ontology construction can effectively help domain experts to construct social security ontology, and also has a considerable reference value for the construction of other domain ontology.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：F239.1;TP391.1

【参考文献】