摘要
提出了利用贝叶斯概率理论进行范例推理,因为在范例库中进行范例检索是范例推理中的一个非常重要的组成部分。该文中主要考虑的是如何利用贝叶斯方法进行范例检索并提出了一个评估两个范例相似程度的匹配函数。在传统的方法中,可以用欧氏距离等方法来评估两个范例的相似性。在实验中,把这两种方法进行了比较,实验结果表明贝叶斯方法不仅可行,而且比用欧氏距离方法更优。
This paper proposes the Bayesian approach for case-based reasoning, because case retrieval in the case base is an essential part of the case-based reasoning, the paper considers how to use the Bayesian approach for case retrival and proposes a case matching function for evaluating the similarities between the cases.In tradictional approaches to the case retrieval,the Euclidean distance is a very important approach.In the experiment, it compares the two approaches.The results suggest that the Bayesian approach is not only feasible,but also performs much better than the Euclidean distance.
出处
《计算机工程》
CAS
CSCD
北大核心
2004年第15期58-59,82,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60273043)
安徽省教育厅资助项目(2002kj004)
关键词
范例推理
贝叶斯方法
范例检索
Case-based reasoning
Bayesian approach
Case retrieval
