| 基于BP模型的优化方法研究及其应用 |
| |
| 引用本文： | 刘桂莲,王福林. 基于BP模型的优化方法研究及其应用[J]. 数学的实践与认识, 2010, 40(6) |
| |
| 作者姓名： | 刘桂莲 王福林 |
| |
| 作者单位： | 东北农业大学工程学院,黑龙江哈尔滨,150030 |
| |
| 基金项目： | 国家“863”专题(2006AA10A310-1) |
| |
| 摘 要： | 首先根据BP算法的基本原理,推导出了BP模型中输出对输入的偏导.然后基于已有数据,利用BP模型进行拟合,保留训练后的BP模型的权值和阈值.在此基础上,利用变步长梯度法实现了BP模型的最优化.本文是对BP模型优化方法的一种新的理论探讨.
|
| 关 键 词： | BP神经网络模型 偏导 最优化 |
The Research on Optimization of BPNN Model and Its Application |
| |
| LIU Gui-lian,WANG Fu-lin. The Research on Optimization of BPNN Model and Its Application[J]. Mathematics in Practice and Theory, 2010, 40(6) |
| |
| Authors: | LIU Gui-lian WANG Fu-lin |
| |
| Affiliation: | LIU Gui-lian,WANG Fu-lin (School of Engineering,Northeast Agricultural University,Harbin 150030) |
| |
| Abstract: | The paper derived out the gradient of the output function based on the principle of Back Propagation Neural Network(BPNN).The proposed ANN model was trained using the data generated from the experiment and preserved the weight and threshold values.The optimization of BPNN model was realized according to the deepest gradient of the output. This is the previous study on the method of the optimization of BPNN model. |
| |
| Keywords: | BPNN modeling gradient optimization |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
