摘要： 基于案例推理的智能推荐系统是大型科学仪器协作共用网的重要组成部分。通过对案例库按网格进行聚类，设计并实现一个基于异构案例库的检索策略。分析案例库网格划分原则及案例聚类规则，论述案例聚类算法及在聚类基础上的案例检索策略。实验结果表明，该方法能够有效地降低案例检索时间，提高案例库的可维护性。

关键词: 基于案例推理, 网格, 相似度, 样本案例

Abstract: The Case-Based Reasoning(CBR) intelligent recommendation system is the important part of scientific instrument shared network. Through clustering the case database in terms of grid, this paper designs a retrieval strategy on the basis of the non-isomorphic case database and also carries it out. It analyzes the principles of case database grid partition and case clustering, and deeply discusses the case clustering algorithm as well as the case retrieval strategy on the basis of clustering. The results of an experiment show that this method can efficiently reduce case retrieval time and enhance case database maintainability.

Key words: Case-Based Reasoning(CBR), grid, similarity, sample case

中图分类号: 

• TP311.13