计算机工程 ›› 2011, Vol. 37 ›› Issue (5): 193-195.doi: 10.3969/j.issn.1000-3428.2011.05.065

• 人工智能及识别技术 • 上一篇    下一篇

基于改进K-means聚类的案例检索策略

乔 丽1,姜慧霖 1,贾世杰2   

  1. (1. 商丘师范学院计算机科学系,河南 商丘 476000;2. 昆明理工大学信息工程与自动化学院,昆明 650051)
  • 出版日期:2011-03-05 发布日期:2012-10-31
  • 作者简介:乔 丽(1980-),女,讲师、硕士,主研方向:智能信息处理,嵌入式技术;姜慧霖,实验师;贾世杰,硕士

Case Retrieval Strategy Based on Improved K-means Clustering

QIAO Li 1, JIANG Hui-lin 1, JIA Shi-jie 2   

  1. (1. Department of Computer Science, Shangqiu Normal College, Shangqiu 476000, China; 2. College of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051, China)
  • Online:2011-03-05 Published:2012-10-31

摘要: 针对目前基于案例推理系统中案例检索存在的问题,根据K-means算法思想,分别设计一个案例聚类算法及案例检索算法。根据K-means算法的不足,对初值选取规则及案例检索算法进行改进。分析基于案例权重的样本案例选取规则,并论述案例聚类算法和检索算法。实验结果表明,该方法能有效提高案例检索效率及案例检索结果的召回率。

关键词: 基于案例推理, 聚类, 案例权重, 相似度

Abstract: Aiming at the case retrieval problems in the Case-Based Reasoning(CBR) system, in the light of the idea of the K-means algorithm, this paper designs a clustering algorithm and a case retrieval algorithm respectively. In terms of the deficiency of the K-means algorithm, this paper improves the selecting rules of initial values as well as the case retrieval algorithm. It analyzes the selecting rules of sample case on the basis of case-weight, and deeply discusses the case clustering algorithm and retrieval algorithm. Experimental results show that this method can efficiently raise the efficiency of case retrieval and enhance the recall rate of the retrieval results.

Key words: Case-Based Reasoning(CBR), clustering, case-weight, similarity

中图分类号: