基于神经网络的预测模型中输入变量的选择  

Input Variables Selection of Forecasting Model Based on Neural Network

作　　者：杨奎河[1] 王宝树[1] 赵玲玲[2] 

机构地区：[1]西安电子科技大学计算机学院,西安710071 [2]河北科技大学信息学院,石家庄050054

出　　处：《计算机科学》2003年第8期139-140,143,共3页Computer Science

摘　　要：It is important to select input variables when the neural network forecasting model is proposed. In this pa-per, by using the autocorrelation function on input variables sets selection for neural network forecasting model, asystemic and scientific method for input variables sets selection is put forward. FFT is adopted to accomplish thespeediness calculation, which enhances the maneuverability of this approach. A forecasting example is given, whoseresult indicates that the method is effective.It is important to select input variables when the neural network forecasting model is proposed. In this paper, by using the autocorrelation function on input variables sets selection for neural network forecasting model, a systemic and scientific method for input variables sets selection is put forward. FFT is adopted to accomplish the speediness calculation, which enhances the maneuverability of this approach. A forecasting example is given, whose result indicates that the method is effective.

关 键 词：神经网络 预测模型 输入变量 数据序列 数学模型 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]