| 基于人工神经网络的多模型综合预报方法 |
| |
| 引用本文： | 路志英,赵智超,郝为,林孔元,刘还珠.基于人工神经网络的多模型综合预报方法[J].计算机应用,2004,24(4):50-51,88. |
| |
| 作者姓名： | 路志英 赵智超 郝为 林孔元 刘还珠 |
| |
| 作者单位： | 1. 天津大学,自动化学院,天津,300072
2. 国家气象中心,北京,100081 |
| |
| 摘 要： | 根据天气系统非线性变化及天气变化受大气多种内外因素综合影响的特点，文中提出了用ANN的前馈网络(BP算法)串入竞争自组织映射网络(SOM网络)方法对同一预报量进行不同结构类型的MOS模型、动力诊断模型和人工智能模型的综合预报。利用这一系统对样本进行了先聚类后训练的预报。结果表明，BP SOM网络实现多模型(异型)综合预报系统具有很好的应用前景。
|
| 关 键 词： | 综合预报 前馈神经网络 SOM神经网络 |
| 文章编号： | 1001-9081(2004)04-0050-02 |
Multi-model Ensemble Forecast Method Based on ANN |
| |
| LU Zhi-ying-,ZHAO Zhi-chao-,HAO Wei-,LIN Kong-yuan-,LIU Huan-zhu-.Multi-model Ensemble Forecast Method Based on ANN[J].journal of Computer Applications,2004,24(4):50-51,88. |
| |
| Authors: | LU Zhi-ying- ZHAO Zhi-chao- HAO Wei- LIN Kong-yuan- LIU Huan-zhu- |
| |
| Affiliation: | LU Zhi-ying-1,ZHAO Zhi-chao-1,HAO Wei-1,LIN Kong-yuan-1,LIU Huan-zhu- |
| |
| Abstract: | Based on the nonlinear characteristic of weather system,an ensemble forecast system of multi-model using multi-layer forward ANN serialized with competition SOM network is presented in this paper. These models are of different types and structures,such as MOS,dynamic diagnosis and AI. With the help of this system,the forecast of clustering followed by training is made on samples. Results show that multi-model (heterogeneous) ensemble forecast realized by BP+SOM network is promising in practice. |
| |
| Keywords: | ensemble forecast feedback neural network SOM neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
