Web工程中基于不变性的元数据发现与聚类
发布时间：2018-12-30 21:37
【摘要】：儿乎所有的Web工程都会使用到元数据,它是Web工程中最常用的配置文件。随着元数据规模的增加,对元数据的维护会花费很多的时间和精力。而目前的编译器不能对元数据不一致所导致的错误进行提示,也不能罗列出元数据和代码之间隐藏的关系。 本文研究并提取了Web工程中的元数据不变性。通过基于框架的元数据不变性发现和框架无关的元数据不变性发现这两种方式来获取元数据不变性,两种元数据不变性的发现方法能覆盖到更多类型的元数据。本文改进了XML的聚类算法并运用到实际的Web工程中,对XML,文件进行分类。基于属性的相似度调整和奇异点的排除使聚类结果更为准确。元数据不变性发现从纵向上发现了XML文件和Java文件的不变性,元数据聚类程序在横向上找出了XML,文件之间的联系。不变性的建立和元数据的分类可以使用户快速地把握项目的架构,发现其中的错误。当用户重构或者增强程序时,元数据不变性会被检查,如果违反了不变性,则对用户进行提示。本文使用了面向对象的设计方法,可以方便地对程序进行扩展。 通过实验,对基于框架的元数据不变性发现和框架无关的元数据不变性发现进行对比,说明了两者的适用情况：验证了元数据不变性发现的有效性。通过与传统聚类方法的对比,体现了改进后的元数据聚类方法的效果：验证了元数据聚类可以有效地对Web工程中的元数据进行分类。
[Abstract]:Almost all Web projects use metadata, which is the most commonly used configuration file in the Web project. With the increase of metadata scale, the maintenance of metadata will take a lot of time and effort. Current compilers cannot prompt errors caused by inconsistent metadata, nor can they list hidden relationships between metadata and code. This paper studies and extracts metadata invariance in Web engineering. Metadata invariance is obtained by frames-based metadata invariance discovery and framework-based metadata invariance discovery. The two metadata invariance discovery methods can cover more types of metadata. This paper improves the clustering algorithm of XML and applies it to the actual Web project to classify the XML, files. The similarity adjustment based on attributes and the elimination of singular points make the clustering results more accurate. Metadata invariance discovery finds the invariance of XML files and Java files vertically, and the metadata clustering program finds out the relationship between XML, files horizontally. The establishment of invariance and the classification of metadata enable users to quickly grasp the structure of the project and find errors in it. Metadata invariance is checked when the user reconstructs or enhances the program, and if the invariance is violated, the user is prompted. In this paper, the object-oriented design method is used to extend the program conveniently. Through experiments, the comparison between frames-based metadata invariance discovery and framework-based metadata invariance discovery is carried out, and the applicability of the two methods is illustrated: the validity of metadata invariance discovery is verified. By comparing with the traditional clustering method, the effect of the improved metadata clustering method is demonstrated. It is verified that the metadata clustering can effectively classify the metadata in Web engineering.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
本文编号：2396187
[Abstract]:Almost all Web projects use metadata, which is the most commonly used configuration file in the Web project. With the increase of metadata scale, the maintenance of metadata will take a lot of time and effort. Current compilers cannot prompt errors caused by inconsistent metadata, nor can they list hidden relationships between metadata and code. This paper studies and extracts metadata invariance in Web engineering. Metadata invariance is obtained by frames-based metadata invariance discovery and framework-based metadata invariance discovery. The two metadata invariance discovery methods can cover more types of metadata. This paper improves the clustering algorithm of XML and applies it to the actual Web project to classify the XML, files. The similarity adjustment based on attributes and the elimination of singular points make the clustering results more accurate. Metadata invariance discovery finds the invariance of XML files and Java files vertically, and the metadata clustering program finds out the relationship between XML, files horizontally. The establishment of invariance and the classification of metadata enable users to quickly grasp the structure of the project and find errors in it. Metadata invariance is checked when the user reconstructs or enhances the program, and if the invariance is violated, the user is prompted. In this paper, the object-oriented design method is used to extend the program conveniently. Through experiments, the comparison between frames-based metadata invariance discovery and framework-based metadata invariance discovery is carried out, and the applicability of the two methods is illustrated: the validity of metadata invariance discovery is verified. By comparing with the traditional clustering method, the effect of the improved metadata clustering method is demonstrated. It is verified that the metadata clustering can effectively classify the metadata in Web engineering.
【学位授予单位】：上海师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.09
【参考文献】
相关期刊论文 前7条
1 夏小彬;严小卫;;XML数据文档及其处理技术探讨[J];计算机科学;2001年08期
2 赵俊岚;XML编程中的DOM与SAX技术[J];计算机工程;2004年24期
3 朴勇;田伟;王秀坤;;一种XML文档结构相似度计算方法[J];控制与决策;2010年04期
4 刘远超;王晓龙;徐志明;关毅;;文档聚类综述[J];中文信息学报;2006年03期
5 查礼;;基于Hadoop的大数据计算技术[J];科研信息化技术与应用;2012年06期
6 陈黎飞;姜青山;王声瑞;;基于层次划分的最佳聚类数确定方法[J];软件学报;2008年01期
7 刘芳,肖铁军;XML应用的基石:XML解析技术[J];计算机工程与设计;2005年10期
，本文编号：2396187
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2396187.html
