案例推理系统中欧式距离检索算法的改进  

Improvement Of Case Retrieval Algorithm Based on Euclidean Distance in CBR System

作　　者：胡小鹏[1] 陆能枝[1] 

机构地区：[1]上海海事大学信息工程学院计算机系,上海200135

出　　处：《计算机与数字工程》2007年第12期35-37,共3页Computer & Digital Engineering

摘　　要：案例检索是基于案例推理系统的中心环节。目前应用最多的欧氏距离检索算法在实际应用中经常出现计算的相似度结果偏离工程实际的情况。利用一种归一化效用函数在最近邻法原理基础上提出一种改进的欧氏距离检索算法。实际工程应用结果证明改进的算法不仅有效而且简单实用。Case retrieval is a focus of the case-based reasoning (CBR) system. The similarity results calculated by the euclidean distance retrieval algorithm, which is widely used in the CBR system, often deviate from the fact in engineering. Based on the Nearest Neighbor approach, a modified euclidean distance retrieval algorithm is put forward by using a normalization utility function. It is shown from application results that the algorithm is simple and practical.

关 键 词：案例推理 最近邻法 欧式距离 相似度算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]