【作者】 刘韬；

【导师】 范菁； 熊丽荣；

【作者基本信息】 浙江工业大学 ， 计算机应用技术， 2008， 硕士

【摘要】 软件复用被视为解决软件危机,提高软件生产效率和质量的现实可行的途径。引入软件复用来实现信用评估系统的构件化开发,可快速构建适应性好、可靠性高的信用评估系统以准确评估企业信用状况,同时有效地减小软件开发成本。如何在大量的软件可复用资产中找到需要的构件是基于构件复用的信用评估软件开发过程中的关键问题。信用构件库提供信用构件的存储、管理、检索以及库的浏览与维护,是支持复用的重要基础设施。信用构件库构件的组织和检索是基于构件的有效表示之上的。本文通过对信用评估领域的分析,提出了信用构件信息的表示方法,并在此基础上,重点讨论了信用构件库采用的检索技术。本文主要在以下几个方面展开研究:1、通过分析构件的刻面分类优势和原则,运用领域工程方法,对信用评估领域分析,提出符合构件提取特征的信用构件刻面分类模式,使信用构件具有了良好的刻面信息表示;2、在信用构件的刻面信息表示基础上,结合树匹配模型和层次编码技术,将信用构件刻面检索转换为字符串的匹配,达到相对的松弛匹配效果,采用向量空间模型VSM(Vector SpaceModel)计算相似度,在保证一定查准率的情况下提高查全率;3、本文运用Lucene全文检索技术实现信用构件库的关键词和属性值的匹配,达到了检索效率的提高,以此适应多层次辅助查询的需求;4、在上述研究的基础上,实现了信用构件库的构件检索子系统,验证了可行性和实用性。本文通过对信用构件库检索技术的讨论,运用了综合检索性能较好的实现方法,对支持基于构件复用的信用评估系统的开发具有一定的促进意义。更多还原

【Abstract】 Software reuse has claimed to be the realistic means to increase the productivity of programmers and improve the quality of developed software. Besides, a well-adapted and reliable credit evaluation system used to evaluate the status of the credit level of enterprises can be developed quickly through the construction of different components, and the cost of the software development can be reduced sharply at the same time. Consequently, how to find the needed components from vast amount of software reusable component library has become the key problem of the development process of credit-evaluation software based on component reuse.The storage, management, retrieval, browses and maintenance of credit component library has been provided, which is the most important infrastructure supporting software reuse. The composition and retrieval of the component library are based on the efficient description of components. Several means to present the evaluation component information has been proposed, based on which the adopted retrieval technology of the evaluation component library has been mainly discussed.This dissertation has done its research on several respects as follows:1. Through analysis of credit evaluation domain based on the advantage and principle of facet classification and domain engineering, a facet classification schema of credit component correspondent to the retrieval characterization has been proposed.2. On the basis of efficient representation of the information of credit components, in conjunction with the technology of tree-match model and hierarchy encoding, the retrieval of component facets is converted to the match between strings to achieve the goal of relaxation matching effect. Through the calculation based of the VSM, the recall has been improved based on the certain level of precision.3. This dissertation has adopted Lucene full text search technology to implement the match of keyword and attribute value for the credit component library. As a result, the efficiency of the retrieval has been greatly improved to satisfy the requirement of multilayer assistant search.4. Based on the above research, this dissertation has implemented the component retrieval system of credit component library, and testifies the feasibility and practicability of the system.Through the discussion of the retrieval technology applied to the credit component library, this dissertation has proposed a good implementation means with good search performance, and facilitate the development of credit evaluation system based on component reuse.更多还原