一种面向大规模构件库的构件检索方法

A COMPONENTS RETRIEVAL METHOD FOR LARGE-SCALE COMPONENT REPOSITORY

针对如何在大规模构件库中改善构件检索性能的问题,提出一种基于功能倒排索引与改进VSM(Vector Space Model)相似度的构件检索方法。应用功能倒排索引能够快速排除在功能上不相关的构件,有效缩小构件检索范围;对VSM相似度算法进行改进,提高构件检索的查准率。与常用检索方法的对比实验表明,该方法有效提高了检索速度,并且检索查全率与查准率也保持在较高水平。

For improving the retrieval performance of large-scale component repositories, a novel approach for components retrieval is proposed based on inverted index of functional facet and the improved algorithm of vector space model （VSM） similarity. The application of the inverted index of functional facet can quick exclude the irrelevant components in function and narrow the range of components retrieval effectively ; the improvement on VSM similarity algorithm also enhances the precision of component retrieval. Comparative experiment on this approach and some common retrieval methods shows that the components retrieval is sped up effectively, the precision and recall ratio is kept on higher level as well.