摘要
为克服BP算法收敛速度慢,易陷入局部极小值等的缺点,提高BP预测精度等性能,提出了变共轭梯度法(VCG)。从激活函数和学习规则两方面修正网络,并对其收敛性作了分析及简要证明。将其应用于我国主要农产品总产量的预测,证实了该算法的有效性。
To improve the prediction accuracy of BP network, the variable conjugate gradient （VCG） algorithm was proposed. The new approach improved BP algorithm from two aspects： activation transfer function and learning rule. Its convergence was analyzed and briefly proved. The variable conjugate gradient algorithm was applied to train a multilayer neural network to predict the total yield of main agricultural products of China, which overcame the slow rate of convergence and poor generalization capability of the traditional BP algorithm.
出处
《计算机应用》
CSCD
北大核心
2006年第11期2765-2768,2772,共5页
journal of Computer Applications
基金
四川省教育厅自然科学重点基金项目(2004A102)
关键词
BP算法
激活函数
学习规则
预测
共轭梯度算法
BP algorithm
activation transfer function
learning nile
prediction
conjugate gradient algorithm
