摘要
为快速有效地实现案例推理,提出了基于多EBP神经网的案例匹配技术,设计了改进的德尔菲法来确定输出样本集.与传统的最邻近法的结果相比,对于复杂系统,该技术能够提高案例匹配的精确性,更适于解决案例属性权重难以确定或存在着耦合与非线性关系的问题.
To implement case reasoning quickly and efficiently,the case retrieving based on multi-EBP neural network (MEBPNN) is proposed in this paper.The improved Delphi approach is designed to determine output sample set.Compared with the result of the conventional nearest neighbor (NN) algorithm, case retrieving precision for complex system can beimproved with MEBPNN,and it is more fitful to resolve some problems in which case attributes weights are difficult to be determined or there exist coupling and non-linear relation among them.
出处
《沈阳工业大学学报》
EI
CAS
2004年第6期711-715,共5页
Journal of Shenyang University of Technology
基金
国家863重大资助项目(2002AA412010)
国家863资助项目(2002AA414610)
关键词
案例推理
动态调度
神经网络
案例匹配
case-based reasoning (CBR)
dynamic scheduling
neural network
case retrieving