• 论文 •    

基于情境树相似性的知识检索技术

郭树行,兰雨晴,金茂忠,高静,李红娟   

  1. 1.北京航空航天大学 软件工程研究所,北京100191;2.内蒙古农业大学 计算机信息与工程学院,内蒙古呼和浩特010018
  • 出版日期:2008-12-15 发布日期:2008-12-25

Knowledge retrieval technology based on situation tree similarity

GUO Shu-hang, LAN Yu-qing, JIN Mao-zhong, GAO Jing, LI Hong-juan   

  1. 1.Software Engineering Institute, Beihang University, Beijing 100191, China; 2.College of Computer Information & Engineering, Inner Mongolia Agricultural University, Hohhot 010018, China
  • Online:2008-12-15 Published:2008-12-25

摘要: 为更好地共享与复用工程知识,提出一种将情境特征作为共享与复用基础的知识检索技术。在该技术中,知识的情境特征被表示为分层和树状模型,每一种工程知识都关联具体的情境树模型。为使当前情境获得适用性高的知识,首先利用情境树相似性进行历史情境的相似性排序,以便在所有历史情境中得到与当前情境最为相似的历史情境,进而逆向根据情境与知识的关联关系,输出选定历史情境所关联的知识。此外,为了改善无序树的树相容匹配算法时间复杂度,提出了一种新的树相容匹配算法KTreeMatch代替传统图论中的无序树匹配算法,从而保证知识检索结果在查全率和查准率上的质量。

关键词: 情境树, 相似性, 知识检索, 树匹配, 知识复用, 知识管理

Abstract: To improve engineering knowledge sharing and reuse, a knowledge retrieval technology was proposed based on knowledge situation characteristics. The situation characteristics of engineering knowledge were expressed as one tree hierarchy model, and each knowledge was associated with concrete situations. In order to retrieve the highly applicable knowledge for the current situation, historical situation similar to current situation could be sorted by its value, and the knowledge could be determined and output based on the association between the highly similar historical situation and the engineering knowledge. In addition, a new tree intersection-matching algorithm, named KTreeMatch was presented upon its models to improve the time complexity of traditional non-ordered tree matching, which ensured the retrieval results with higher recalling and precision.

Key words: situation tree, similarity, knowledge retrieval, tree matching, knowledge reuse, knowledge management

中图分类号: