构件的标识表示与检索方法研究  

Study on Tags Representation of Components and Tags Based Components Retrieval

作　　者：张雷[1] 陈立潮[1] 潘理虎[1,2] 闫惠敏[2] 张英俊[1] 

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024 [2]中国科学院地理科学与资源研究所,北京100101

出　　处：《小型微型计算机系统》2013年第5期1076-1079,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(41071344)资助;山西省自然科学基金项目(2009011022-1)资助;太原科技大学研究生创新项目(20111025)资助

摘　　要：针对如何在大规模构件库中改善检索性能的问题,提出构件的标识表示,并设计了基于标识的构件检索方法.标识由领域术语、刻面术语以及高权重词等代表构件主要信息的词组成.使用标识表示构件可以获得比刻面分类更多的有效信息并能够校正刻面分类产生的主观影响,有利于检索准确率的提高.为提高检索速度,设计了功能倒排索引与标识索引组成的双向组合索引.最后与常用检索方法进行了对比实验,结果表明基于标识的构件检索具有较高的检索效率以及良好的综合检索效果,能够较好的满足大规模构件库的检索要求.For improving the retrieval performance of large-scale component repositories, a tag representation of components is proposed, and based on the tag representation, a method of tag retrieval is designed. Tags consist of terms which express the main information on components such as domain terms, facet terms, high weight terms and so on. Since contained more information than facets, tag representation can effectively overcome the subjective effects of faceted classification and help to improve the precision of component retrieval. In order to improve the retrieval speed, a combined index of functional facet and tags is designed. At last, the experiment results show that tagbased retrieval has higher retrieval efficiency and better performance than common retrieval methods, and the requirements of retrieval in large-scale component repository can be met effectively with this method.

关 键 词：软件构件 标识表示 标识检索 组合索引 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]