摘要
采用基于该神经网络技术的案例推理系统,使用交叉覆盖算法,可以有效地缩减案例的检索时间、减少案例适应性修改、提高推理效率。实验表明该系统易于设计构建,极大地提升了CBR在实际中的应用能力。
This paper presents a CBR system based on multi-layered feedforward neural network and its alternative-covering algorithm, which can greatly decrease the time taken to perform case retrieval and matching. The experimental results indicate that the integrated technology can efficiently enhance the system performance, especially for the large-scale case-based reasoning, which can facilitate CBR system design and promote the capacity of applying CBR to real-world problem-solving.
出处
《计算机工程》
CAS
CSCD
北大核心
2006年第7期188-190,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60273043)
安徽省自然科学基金资助项目(050460402)
关键词
CBR
集成系统
前馈神经网络
交叉覆盖算法
Case-based reasoning
Hybrid system
Feedforward neural network
Alternative-covering algorithm
作者简介
李建洋（1968-），男，副教授、博士生，主研方向：机器学习，神经网络，智能决策支持系统； E-mail：lijianyang@sina.com
郑汉垣，副教授；
刘慧婷，博士生
