基于等级策略的构件聚类研究  

Research of Component Clustering Based on the Ranking Strategy

作　　者：王春红[1] 任姚鹏[1] 徐学洲[2] 

机构地区：[1]运城学院计算机科学与技术系,山西运城044000 [2]西安电子科技大学软件工程研究所,陕西西安710071

出　　处：《计算机工程与科学》2010年第5期118-120,158,共4页Computer Engineering & Science

基　　金：山西高校科技研究开发项目(20091150)

摘　　要：构件库系统需解决两个核心问题:构件的分类表示与构件的检索,其中对构件的合理分类是实现构件高效检索的基础和前提。针对目前常用的刻面分类表示法存在的人为主观因素,采用了刻面分类表示与正文检索相结合的方法对构件进行描述,并依据用户的需求对刻面分类方案中的主刻面赋予不同的等级。提出了一种基于等级策略的构件聚类方法对构件进行聚类分析,并通过实验验证了该方法是有效的,实现了对构件更合理的分类,提高了构件检索的效率和准确性。The Component Library System needs to solve two core problems, which are component classification repre- sentation and component retrieval. Meanwhile, to implement component efficient retrieval is based on the component rational classification. Aiming at the subjective factors of the faceted classification representation, the component is described by the way which combines the text retrieval with the faceted classification representation, and according to the user require- ments, different ranks are given to the main facets of the faceted classification schema. Then, a component clustering meth- od based ranking strategy is proposed. The experiments prove the method is effective, which implements a more rational classification of the components, and improves the efficiency and accuracy of component retrieval.

关 键 词：等级策略 构件聚类 刻面分类 正文检索 向量空间模型 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]