摘要
为了更好的满足嵌入式应用领域和实时性环境的要求,在资源分配网络(RAN)的基础上提出了一种改进型径向基函数(RBF)神经网络在线学习算法。在网络参数调整过程中引入了分级学习率因子,根据理论输出和网络输出误差绝对值的大小选择不同的学习率因子参与学习过程。在VC6.0编程环境中进行的软件仿真试验表明:相对于传统的RBF神经网络在线学习算法,改进型RBF神经网络在线学习算法在不增加网络规模的情况下可以进一步减小输出误差。
In order to meet the requirements of embedded applications and real-time environment better, an improved on-line learning algorithm of Radial Basis Function(RBF)neural network based on resource allocation network(RAN)is proposed.A hierarchical learning rate factor is introduced into this algorithm in the network parameter adjustment process.Different learning rate factors are selected to take part in the learning process according to the absolute value of the error between theoretical output and network output.VC6.0software simulation experiments show that compared with the traditional RBF neural network online learning algorithm,the improved RBF neural network on-line learning algorithm can further reduce the output error without increasing the size of the network.
出处
《常州大学学报（自然科学版）》
CAS
2014年第1期52-56,共5页
Journal of Changzhou University：Natural Science Edition
