| 摘要: |
| 根据地下水及其影响因素之间存在的映射关系，在BP网络模型的基础上，提出一种Levenberg-Marquart优化神经网络算法，并用于地下水位的预测．与传统的BP算法相比较，该算法的预测精度较高，计算结果稳定性好，收敛速度快． |
| 关键词: 神经网络 Levenberg—Marquart算法 地下水位 预报 |
| DOI： |
| 分类号:P641.2 TP183 |
| 基金项目: |
|
| Prediction of groundwater level by optimized neural network algorithm |
|
CHANG Liang,XIE Jian-cang
|
| Abstract: |
| Based on the reflection relationship between the groundwater level and its influence factors and BP neural network model, a Levenberg-Marquart optimization algorithm of BP neural network is proposed and used for prediction of groundwater level. According to the comparison between the prediction results and the conventional BP algorithm, it can be seen that this algorithm has advantages of high prediction accuracy, fine stability and fast convergence speed. |
| Key words: neural network Levenberg-Marquart algorithm groundwater level prediction |
