Feedback GMDH-Type Neural Network Algorithm Using Prediction Error Criterion Defined as AIC
Feedback GMDH-Type Neural Network Algorithm Using Prediction Error Criterion Defined as AIC
复制标题
DOI：
10.1007/978-3-642-29920-9_32
复制
发表时间：
2012-05
影响因子：
2.3
通讯作者：
T. Kondo;J. Ueno;S. Takao
中科院分区：
文献类型：
--
作者：
T. Kondo;J. Ueno;S. Takao
提出了一种基于AIC预测误差准则的反馈群数据处理(GMDH)型神经网络算法。该算法从Sigmoid函数型神经网络、径向基函数(RBF)神经网络和多项式神经网络三种网络结构中自动选择最优的神经网络结构。此外，为了最小化定义为Akaike信息准则(AIC)的预测误差准则，自动选择反馈回路数目、隐含层神经元数目和有用输入变量等结构参数。反馈GMDH型神经网络有一个反馈环，通过反馈环的计算使神经网络的复杂度逐渐增加，以适应非线性系统的复杂性。将该算法应用于复杂非线性系统的辨识问题。
In this study, a feedback Group Method of Data Handling (GMDH)-type neural network algorithm using prediction error criterion defined as AIC, is proposed. In this algorithm, the optimum neural network architecture is automatically selected from three types of neural network architectures such as sigmoid function type neural network, radial basis function (RBF) type neural network and polynomial type neural network. Furthermore, the structural parameters such as the number of feedback loops, the number of neurons in the hidden layers and useful input variables are automatically selected so as to minimize the prediction error criterion defined as Akaike’s Information Criterion (AIC). Feedback GMDH-type neural network has a feedback loop and the complexity of the neural network increases gradually using feedback loop calculations so as to fit the complexity of the nonlinear system. This algorithm is applied to identification problem of the complex nonlinear system.