摘要
案例匹配在基于案例的推理(CBR)中非常关键[3],直接影响到CBR系统的有效性和高效性.在某些CBR的应用领域中,案例比较复杂,案例匹配的实现也比较困难,匹配结果的准确性和匹配速度都难以保证.本文提出的基于智能聚类的复杂案例匹配方法,为提高匹配结果的准确性和匹配速度提供了一条良好的途径.
Processing complex cases is difficult in CBR. A complex case retrieval method based on intelligent clustering is presented in this paper. An object-oriented aggregate representation pattern is designed to represent complex cases. Hierarchy and fuzzy logic is used to measure similarity between complex cases to improve the accuracy of match results. In order to increase the speed of case retrieval, SOM neural network is applied for clustering of complex cases. Generally, several SOM networks are needed for different parts of complex cases to complete the overall clustering. Then the match process is simplified to find matching cases from the previous cases the same kind as the new problem.
出处
《计算机科学》
CSCD
北大核心
2002年第4期75-78,共4页
Computer Science
关键词
模糊集
模糊数学
隶属函数
智能聚类
案例匹配
神经网络
推理
Case-Based Reasoning, Complex case retrieval, Object-oriented, Fuzzy logic, Self-Organizing Map
