一种改进的 RBF 神经网络对县级政府编制预测

A Prediction for the Preparation of  County Government Based on  Improved RBF Neural Networks

  • 摘要: 针对未来政府编制总量的预测,提出了一种改进的 RBF 网络算法,通过引入 GCV 准则进一步优化宽度参数 σ;同时,对 RBF 网络进行子网络化处理以优化网络性能.实验结果表明,采用改进的 RBF 网络模型能够进一步提高网络的拟合精度,比传统的编制总量预测方法误差更小,预测精确度更高.

     

    Abstract: To predict the future total amount of the governmental preparation, an improved algorithm of RBF network was brought up by introducing GCV criterion to optimize the width parameter 〖WTBX〗σ〖WTBZ〗 and by dividing the RBF network into sub-network to optimize the performance of the network. The experimental results showed that the RBF network model based on the improved RBF network algorithm can provide smaller errors and better results than the traditional methods of total account of staffing prediction in practice, and the prediction accuracy was higher. 

     

/

返回文章
返回