摘要
针对一般模糊极小极大(generalfuzzymin max,GFMM)神经网络不能够完全无师聚类和自适应在线学习的问题,提出了一种无师训练的一般模糊极小极大(generalfuzzymin max,GFMM)人工神经网络。它继承了GFMM网络的优点,可以输入n维模糊量,尤其是新增加了无师学习的功能,弥补了GFMM网络不能自适应在线学习新类的缺陷。实验测试结果与分析表明,该网络在自动目标识别的实际应用中具有广泛的适用性。
An unsupervised general fuzzy min-max(GFMM) artificial neural network is proposed. This network inherits the merit of the general fuzzy min-max network which uses the fuzzy input vectors. Because of the newly added unsupervised learning function, it counteracts the weakness which makes the general fuzzy min-max network be incapable of learning any new pattern class. The results and analyses of the experimental testing indicate that this network will find a wide application in the automatic target recognition in the future.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2004年第10期1503-1505,1536,共4页
Systems Engineering and Electronics
关键词
一般模糊极小极大神经网络
无师训练
自动目标识别
general fuzzy min-max neural network
unsupervised training
automatic target recognition
