摘要
东营市短期气候预测系统使用逐步回归、最优子集回归、BP人工神经元网络和多因子EOF迭代4种因子预报模型,以及3个以均生函数为基函数的多步预测模型,并通过集成得出预报结论。因子普查时,采用滚动式相关分析方法,保证了入选因子的稳定性。
We introduce the researching method and the application circumstances of the short - range climate prediction system in Dongying city. The system includes stepwise regression,optimal subset regression ,BP -artificial neural network and multi - factors EOF iteration models and three multi - steps models based on mean generating function, then we get the conclusion through integration. The rolling- correlative analysis method make the factors more steadyly.
出处
《河南气象》
2006年第1期23-24,共2页
Meteorology Journal of Henan
关键词
滚动式相关分析
预测模型
集成预报
业务应用
Rolling correlative analysis
Forecast model
Integrated prediction
Operational application