摘要
本体映射的核心在于语义相似度算法,单一的概念相似度计算方法往往不利于提高相似度的精度。本文针对机械零部件领域本体(MPO)提出一种基于本体加权树的语义相似度算法OWSTS,利用MPO提取领域知识文档标题信息中的核心概念,并结合OWSTS算法来确定文档信息与查询式间的语义关联程度。该方法在GB_MPO智能信息检索系统中得到较好的应用。实验表明,该方法与基于TF*IDF的信息检索方法相比,检索性能有较大提高。
The semantic similarity algorithm is the key to ontology mapping. The past experiences tell us that single concept similarity algorithms are not in favor of improving the precision of retrieving the results. In the paper, according to MPO, we propose a new semantic similarity algorithm based on the weighted tree of the domain ontology. With the help of MPO, we can extract key concepts from the title of document, and then apply OWSTS to evaluating the association degree between documents and retrieving expressions. This approach has been well applied in the intelligent retrieving system, GB_MPO. The results of experiments show that the performance of the OWSTS algorithm is obviously better than that of TF ＊ IDF.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第9期77-79,144,共4页
Computer Engineering & Science
基金
国家自然科学基金资助项目(50175103)
关键词
领域本体
加权树
语义相似度
智能检索
MPO
domain ontology
weighted tree
semantic similarity
intelligent retrieval
MPO
