基于SVM的RBF网络在农机总动力预测中的应用
|
摘要
径向基神经网络是一种单隐层的三层前向网络,具有结构简洁、学习速度快等优点。为此,分析了径向基神经网络采用传统聚类方法确定基函数中心存在的问题,提出了一种基于支持向量机聚类确定径向基网络基函数中心的方法。该方法以最大间隔原理和结构风险最小化原则为前提,利用核方法把输入空间的样本映射到高维特征空间完成聚类工作来确定基函数的数量。采用改进的方法训练的径向基神经网络对黑龙江省农机总动力进行非线性时间序列预测,结果表明:改进的网络在确定网络结构、学习速度和提高网络预测精度方面都有较好的效果。
Radial basis function neural network is a single hidden layer three-layer forward network,which has the advantages of simple structure,fast learning speed and so on. This paper analyzes the problem of using radial basis neural network to determine the existence of basis function center by using traditional clustering method,and proposes a method based on support vector machine clustering to determine the basis function center of radial basis network. This method uses maximum interval principle and structural risk. The principle of minimization is premised. The kernel method is used to map the input space samples to the high-dimensional feature space to complete the clustering work to determine the number of basis functions.Radial basis function neural network with improved training method is used to predict the total power of farm machinery in Heilongjiang Province. The experimental results show that the improved network has a good effect in determining the network structure,learning speed and improving the network prediction accuracy.
Radial basis function neural network is a single hidden layer three-layer forward network,which has the advantages of simple structure,fast learning speed and so on. This paper analyzes the problem of using radial basis neural network to determine the existence of basis function center by using traditional clustering method,and proposes a method based on support vector machine clustering to determine the basis function center of radial basis network. This method uses maximum interval principle and structural risk. The principle of minimization is premised. The kernel method is used to map the input space samples to the high-dimensional feature space to complete the clustering work to determine the number of basis functions.Radial basis function neural network with improved training method is used to predict the total power of farm machinery in Heilongjiang Province. The experimental results show that the improved network has a good effect in determining the network structure,learning speed and improving the network prediction accuracy.
引文
[1]王吉权,王福林,邱立春.基于BP神经网络的农机总动力预测[J].农业机械学报,2011,42(12):121-126.
[2]马海志,王福林,王慧鹏,等.基于改进BP神经网络的黑龙江农机总动力预测[J].农机化研究,2016,38(2):22-30.
[3]何志连,王福林,董慧英,等.BP神经网络最佳停止法对农机总动力的预测[J].农机化研究,2017,39(2):1-5.
[4] Broomhead DS,Lowe D.Multi-variable functional interpolation and adaptive networks[J]. Complex System,1988(2):321-355.
[5] Moody JE,Darken CJ. Fast learning in networks of locallytuned processing units[J]. Neural Computation,1989(2):281-294.
[6]潘琪.径向基神经网络基函数中心确定方法改进研究[D].哈尔滨:东北农业大学,2017.
[7]王福林.农业系统工程[M].北京:中国农业出版社,2017.
[8]李航.统计学习方法[M].北京:清华大学出版社,2012.
[9]王吉权,王福林,邱立春.基于BP神经网络的农机总动力预测[J].农业机械学报,2011,42(12):121-126.
[2]马海志,王福林,王慧鹏,等.基于改进BP神经网络的黑龙江农机总动力预测[J].农机化研究,2016,38(2):22-30.
[3]何志连,王福林,董慧英,等.BP神经网络最佳停止法对农机总动力的预测[J].农机化研究,2017,39(2):1-5.
[4] Broomhead DS,Lowe D.Multi-variable functional interpolation and adaptive networks[J]. Complex System,1988(2):321-355.
[5] Moody JE,Darken CJ. Fast learning in networks of locallytuned processing units[J]. Neural Computation,1989(2):281-294.
[6]潘琪.径向基神经网络基函数中心确定方法改进研究[D].哈尔滨:东北农业大学,2017.
[7]王福林.农业系统工程[M].北京:中国农业出版社,2017.
[8]李航.统计学习方法[M].北京:清华大学出版社,2012.
[9]王吉权,王福林,邱立春.基于BP神经网络的农机总动力预测[J].农业机械学报,2011,42(12):121-126.