使用本体推理技术获取ER模型

发布时间：2018-05-16 02:23

本文选题：本体推理 + 数据库设计　；参考：《吉林大学》2017年硕士论文

【摘要】：数据库自诞生以来一直影响着人们的日常生活,方便了数据的存储和共享。在当前数据膨胀的时代,如何有效地管理各种各样的信息资源,是信息系统的核心工作,也是我们不断研究的重要方向。在现实应用中,经常会由于各种实际问题给数据库设计带来了很多困难。例如:数据库设计者的领域知识缺乏,用户的需求表达不明确,用户和数据库设计者在沟通表达上存在理解的差异等。实体关系模型(ER模型)作为概念模型的一种工具能够将现实生活中的信息抽象为图形化的表示形式,进而直观的表达概念和概念之间的关系。ER模型可以有效地解决数据库设计过程中存在的问题,因此设计一个完善的实体关系模型是数据库设计的核心工作。随着网络技术的不断发展,领域专家将领域知识上传到网络上供大家使用,人们借助于网络的开放性和共享性实现了信息的互联。领域本体不仅能够形式化的表示领域知识,而且能够进行推理,人们可以通过本体推理技术和规则从领域本体中获取希望得到的信息。本文利用网络上丰富的本体资源和领域知识作为数据来源,提出了一种新的构建数据库的方法。结合用户需求,通过自定义的算法和规则,使用本体推理技术获取符合用户需求的实体关系模型。本文将详细地介绍如何使用本体推理技术从本体知识中获取ER模型,并通过一个通用的大学本体对本文提出的方法进行了验证。此外,为了验证本文提出的方法具有普遍适用性,随机选取了多个领域的不同本体进行实验,进一步验证了我们提出的方法是可行有效的。本文使用本体推理技术,不仅能够获取显性的领域知识而且能够蕴含隐含的知识,为ER模型的构建提供了丰富的数据来源。借助于本文提出的方法,一方面很好的弥补了设计者对于领域知识缺乏的不足;另一方面始终以用户需求为出发点,能够保证最终构建的ER模型符合用户的实际需求,减少了设计者和用户反复沟通确认的繁琐流程。本文研究的主要内容有以下几方面:(1)根据用户给出的需求术语,利用本体查询工具找到网络上相应的本体资源作为元数据(metabase),调用推理机获取本体中的概念层次关系,然后对元数据进行预处理。(2)根据预处理得到的概念集合,通过推荐实体集的操作模块获取最终用于构建ER模型的实体集。主要包括三个处理过程:首先对预处理得到的概念集合进行扩展,使用本文提出的算法结合本体推理技术,获取初始实体集(initial Entities简写为init Entities),然后根据自定义的删除规则对初始实体集中的部分概念进行删除,得到候选实体集(candidate Entities,简写为candi Entities),最后通过自定义的修改规则对候选实体集中孤立的概念进行修改,得到最终用于构建ER模型的实体集(Entities)。(3)使用实体集(Entities)和本体推理技术获取实体集之间的关系集。为了使最终的ER模型简洁明了,根据自定义的修改规则对存在包含关系的关系集进行修改,获取最终用于构建ER模型的关系集。(4)使用本体推理技术,获取关系类型和属性,并对子概念(sub Class)继承父概念(super Class)属性的情况加以介绍。通过获取关系类型和属性,对只有实体集和关系集的简易ER模型进行了丰富,在保证满足用户需求的基础上,使得ER模型表达的语义更加丰富。
[Abstract]:Since the birth of the database has been affecting people's daily life, it is convenient for the storage and sharing of data. In the current time of data expansion, how to manage all kinds of information resources effectively is the core work of the information system, and it is also an important direction for our continuous research. In practical applications, it is often due to various practical problems. There are many difficulties in database design, such as lack of domain knowledge of database designers, uncertainty of user requirements expression, and differences in understanding between users and database designers in communication expression. The entity relationship model (ER model) can abstract the information of real life into graphics as a tool for conceptual model. The expression form, and then intuitively express the relationship between concepts and concepts,.ER model can effectively solve the problems existing in the process of database design, so designing a perfect entity relationship model is the core work of database design. With the continuous development of network technology, domain experts upload domain knowledge to the network. For everyone to use, people have realized the interconnection of information with the help of the openness and sharing of the network. Domain ontology can not only form a formal representation of domain knowledge, but also can carry out reasoning. People can obtain the desired information from the domain ontology through the ontology reasoning technology and rules. This paper uses the rich ontology resources on the network. And domain knowledge as a data source, a new method of building a database is proposed. Based on user requirements and custom algorithms and rules, ontology reasoning is used to obtain the entity relationship model which meets the requirements of users. This paper will introduce how to use ontology technology to obtain ER model from ontology knowledge. A general university ontology has verified the method proposed in this paper. In addition, in order to verify the universal applicability of the method proposed in this paper, we randomly select the different ontology of various fields to carry out experiments, and further verify that the method proposed is feasible and effective. This paper can not only obtain dominance using ontology reasoning technology. The domain knowledge and the implicit knowledge can provide a rich source of data for the construction of the ER model. With the help of the method proposed in this paper, it is good to make up for the shortage of the designer for the lack of domain knowledge. On the other hand, we always take the user needs as the starting point, and can guarantee the final construction of the ER model to conform to the user. The main contents of this paper are as follows: (1) according to the requirement terms given by the user, the ontology resources on the network are found by the ontology query tool as metadata (metabase), and the reasoning machine is called to obtain the concept hierarchy in the ontology. Then the metadata is preprocessed. (2) according to the concept set obtained by preprocessing, the entity set which is finally used to build the ER model is obtained by the operation module of the recommended entity set. It mainly includes three processing processes: first, it extends the concept set obtained by preprocessing, and uses the algorithm proposed in this paper to combine the ontology reasoning technology to obtain the initial stage. The entity set (initial Entities is init Entities), and then deletes part concepts of the initial entity set according to the custom deletion rules, and gets the candidate entity set (candidate Entities, the short term as candi Entities). Finally, the concept of the candidate entity is isolated by the custom modification rules, and finally the concept of the candidate entity is modified. The entity set (Entities) is used to build the ER model. (3) using the entity set (Entities) and ontology reasoning technology to obtain the relationship set between the entity sets. In order to make the final ER model concise and clear, according to the custom modification rules, the relation set of the existence relation is modified, and the relationship set which is finally used to build the ER model is obtained. (4) the use of this model is used. The body reasoning technology obtains relation type and attribute, and introduces the case of inheriting parent concept (super Class) attribute of sub Class. By obtaining relation type and attribute, it enriches the simple ER model with only entity set and relation set, and makes the semantics of ER model more expressed on the basis of guaranteeing the requirement of the user. Rich.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】