| 基于主分量神经网络的降水集成预报方法研究 |
| |
| 引用本文： | 农孟松,黄海洪,孙崇智,郑凤琴,陈伟斌.基于主分量神经网络的降水集成预报方法研究[J].气象,2011,37(3):352-355. |
| |
| 作者姓名： | 农孟松 黄海洪 孙崇智 郑凤琴 陈伟斌 |
| |
| 作者单位： | 1. 南京信息工程大学大气科学学院,南京,210044;广西气象台,南宁,530022
2. 广西气象台,南宁,530022 |
| |
| 基金项目： | 中国气象局新技术推广项目预报员专项（CMATG2008Y07）和广西科技厅攻关项目(桂科攻0993002-1和0816006-9)共同资助 |
| |
| 摘 要： | 运用人工神经网络与主分量分析(PCA)相结合的方法,对同一降水预报量的各种数值预报产品进行集成预报研究.结果表明:主分量人工神经网络方法所构造的集成预报模型,不仅对历史样本的拟合精度好于个各子预报产品,独立样本的实验预报结果也显示出更好的预报准确率及稳定性.业务应用前景良好.
|
| 关 键 词： | 主分量 神经网络 集成预报 |
| 收稿时间： | 1/7/2010 12:00:00 AM |
| 修稿时间： | 4/7/2010 12:00:00 AM |
A Neural Network Model Based on Principal Component Analysis for Ensemble Precipitation Prediction |
| |
| NONG Mengsong,HUANG Haihong,SUN Chongzhi,ZHENG Fengqin and CHEN Weibin.A Neural Network Model Based on Principal Component Analysis for Ensemble Precipitation Prediction[J].Meteorological Monthly,2011,37(3):352-355. |
| |
| Authors: | NONG Mengsong HUANG Haihong SUN Chongzhi ZHENG Fengqin and CHEN Weibin |
| |
| Institution: | 1 School of Atmospheric Sciences, Nanjing University of Information Science and Technology, Nanjing 210044,2 Meteorological Observatory of Guangxi, Nanning 530022;2 Meteorological Observatory of Guangxi, Nanning 530022;2 Meteorological Observatory of Guangxi, Nanning 530022;2 Meteorological Observatory of Guangxi, Nanning 530022;2 Meteorological Observatory of Guangxi, Nanning 530022 |
| |
| Abstract: | Using the method of artificial neural network and principal component analysis (PCA) to study a variety of numerical forecast products for the same precipitation forecast. The results show that the fitting accuracy of the principal component analysis artificial neural network ensemble model is better than each sub product, and the experimental results of the independent sample also show its better prediction accuracy and stability. The model is a good prospect for operational applications. |
| |
| Keywords: | principal component analysis neural network ensemble prediction |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《气象》浏览原始摘要信息 |
| 点击此处可从《气象》下载免费的PDF全文 |
|
