摘要
提出了一种人工免疫学习算法,该算法具有识别多样性、自我调节功能等特点。将该算法用于RBF神经网络隐层中心点数量和位置的调整,网络的输入数据作为抗原,抗体(RBF中心)作为抗原的内影像,输出采用最小二乘法。仿真结果表明,该混合算法具有较强的泛化能力,而且精度高,效果好。
This paper presents a learning algorithm b ased on artificial immune which has the characteristics of recognition variable and self-adjustment.It is used for the choice of the number and position for t he center of the hide layer of RBF neural network.The input of the network is the antigen with its internal image as the antibody.The least square algorithm is adopted for the weight training.The simulation study shows that the algori thm is of good generalization ability and high precision.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第1期39-40,46,共3页
Computer Engineering and Applications
基金
高等学校博士学科点专项科研基金资助(编号:20020008004)
关键词
人工免疫
RBF神经网络
中心点
混合算法
artificial immune,radial basis function neural network,center,hybrid algorithm
