Enhancing UML Class Diagram Abstraction with Knowledge Graph
Enhancing UML Class Diagram Abstraction with Knowledge Graph
复制标题
使用知识图增强 UML 类图抽象
DOI：
10.1007/978-3-319-46257-8_65
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Chuan Zhu
中科院分区：
文献类型：
--
作者：
Liang Huang;Yucong Duan;Xiaobing Sun;Zhaoxin Lin;Chuan Zhu
模型驱动工程（Model-Driven Engineering， MDE）通过从模型生成代码，减轻了认知的复杂性和花费在软件开发上的工作量。在MDE中，模型要准确、精炼、可靠、高效。类图是实际系统的结构抽象，通常用于软件设计。一个设计得更好的类图可以带来一个更好的系统。本文提出了一种基于知识图的改进类图的方法。为了方便信息导入，我们将知识图作为媒介层，提出了在类图和知识图之间双向映射数据、信息和知识的方法。基于添加的知识源，设计层次聚类算法对类图进行抽象，最终自动生成抽象的类图。
Model-Driven Engineering (MDE) alleviates the cognitive complexity and effort spent on software development by generating codes from models. In MDE, models should be accurate, refined, reliable and efficient. Class diagram is a structural abstraction of a real system and usually used in software design. A better designed class diagram could lead to a better system. In this paper, we proposed a knowledge graph based method to improve class diagrams. We took knowledge graph as the media layer for easier information introduction, and proposed methods to map data, information and knowledge between class diagrams and knowledge graphs bidirectionally. Based on the added knowledge source, we designed hierarchical clustering algorithm to abstract the class diagram, and finally we generated abstracted class diagrams automatically.