摘要
提出了一种电离层参数短期预报的综合模型,可综合使用几种预报方法进行预报,合理确定不同方法参与预报的权重,能够充分发挥各种预报方法的优势,确保达到最佳的预报性能。作为例子,以自相关方法、相似日方法和神经网络方法作为模型中的方法输入,利用中国9个电离层观测台站一个太阳活动周期(1977~1987年)的f0F2数据,对提前1~24h预报的精度进行了测试。结果表明:综合模型的预报精度高于任一单一模型,使用综合预报模型可以综合考虑各模型的优势,从而得到更好的预报效果。
Multiple methods are developed for ionospheric short-term forecasting home and abroad, so that various results can be got. An integrated model of ionospheric short-term forecasting is presented which can integratedly use various methods for forecasting, properly determine the weights of the methods and make full use of their advantages, so that the optimal forecasting performance can be achieved. For examples, autocorrelation analyzed method, similar day method and neural networks method are used as inputs, a test of the integrated model for 1- 24h forecast of nine ionospheric observation stations in China during a solar activity cycle(1977-1987) shows that the integrated model is better than each single method, and more accurate results can be obtained.
出处
《电波科学学报》
EI
CSCD
北大核心
2010年第3期491-498,共8页
Chinese Journal of Radio Science
基金
国家基础研究项目
国家自然科学基金项目(编号60771049)