| 多模式集成的RBF神经网络天气预报 |
| |
| 引用本文： | 熊聪聪,潘 璇,赵 奇,吴振玲.多模式集成的RBF神经网络天气预报[J].天津科技大学学报,2014(1):75-78. |
| |
| 作者姓名： | 熊聪聪 潘 璇 赵 奇 吴振玲 |
| |
| 作者单位： | [1] 天津科技大学计算机科学与信息工程学院,天津300222 [2] 天津市气象局,天津300074 |
| |
| 基金项目： | 中国气象局气象关键技术集成与应用项目(CAMGJ2012M04);天津市科技型中小企业技术创新资金项目(12ZXCXGX33500) |
| |
| 摘 要： | 针对复杂庞大的多模式数值预报数据,提出一种径向基函数(RBF)神经网络集成天气预报模型.根据天津市预报站点采用的WRF模式、RUC模式等数值预报数据的特点,将多种单模式数据作为RBF神经网络输入,网络输出为集成预报结果.实验表明:RBF神经网络集成预报模型降低了单模式预报误差,更加贴近了真实数据,并且在稳定性和实效性方面均有良好表现.
|
| 关 键 词： | 集成预报 多模式 神经网络 径向基函数 |
RBF Neural Network for Weather Forecast Based on Multi-model Integration |
| |
| XIONG Congcong,PAN Xuan,ZHAO Qi,WU Zhenling.RBF Neural Network for Weather Forecast Based on Multi-model Integration[J].Journal of Tianjin University of Science & Technology,2014(1):75-78. |
| |
| Authors: | XIONG Congcong PAN Xuan ZHAO Qi WU Zhenling |
| |
| Institution: | XIONG Congcong, PAN Xuan, ZHAO Qi, WU Zhenling |
| |
| Abstract: | An integrated forecast model based on radial basis function（RBF）neural network was proposed for large com-plex multi-model numerical forecasting data. According to the characteristics of the numerical model forecast data of WRF model and RUC model used in Tianjin,numerical data of several models were chosen as the input of the RBF neural network,and the output is the integrated result. Experiments of temperature integration show that the RBF neural network integration method can reduce the error of the single model. The integrated result does good work in simulating real data. The method also has stability and effectiveness. |
| |
| Keywords: | integrated forecast multi-model neural network radial basis function |
| 本文献已被 CNKI 维普 等数据库收录！ |
