基于多层前馈神经网络的案例推理系统  

Case-based Reasonor Based on Multi-layered Feedforward Neural Network

作　　者：李建洋[1] 郑汉垣[2] 刘慧婷[1] 

机构地区：[1]合肥工业大学计算机网络所,合肥230009 [2]龙岩学院计算机科学系,龙岩364000

出　　处：《计算机工程》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)

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.

关 键 词：CBR 集成系统 前馈神经网络 交叉覆盖算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]