摘要
ＢＰ网络是迄今为止应用最广泛的一种神经网络，但这种算法也存在着收敛速度慢、容易陷入局部极小点等问题．本文在标准ＢＰ算法的基础上提出一种改进ＢＰ算法，称之为自适应ＢＰ算法．这种自适应ＢＰ算法采用模糊规则动态调整学习参数，并且能在学习过程中和学习完成后通过隐节点调整算法优化网络结构。
BP algorithm is probably the most popular and widely used neural networks. However, the algorithm has a low convergence rate and is easily trapped in local minima. In this paper, we present a new algorithm based on the standard BP algorithm, called adaptive BP algorithm. The adaptive BP algorithm uses a fuzzy controller to adjust the learning parameters dynamically and is able to optimize the network structure by adjusting the number of hidden nodes during and after training procedure. The adaptive BP algorithm has been tested to have higher convergence rate and perform better in generalization.
出处
《系统工程学报》
CSCD
1997年第1期55-62,共8页
Journal of Systems Engineering
基金
国家自然科学基金
关键词
神经网络
BP算法
模糊推理
neural network, BP algorithm, fuzzy reasoning
