摘要
本文在传统时间序列分析建模方法的基础上，提出了用ＡＲ 模型的新的完整而又简单的辨识方法。参数估计采用计算简单的序贯最小二乘解法，并提出了决定模型阶的Ｆ检验判决法，从而可得节省参数模型并用于预测。
On the basis of traditional time series analysis and modeling methods, the thesis puts forward a new complete and simply identification method by using AR model. Parameter estimation of this method adopts sequential Least Square method, the thesis also puts forward how to determine model order by using F test judger, so a parameter saving model is obtained and can be used for forecasting.
出处
《测绘学报》
EI
CSCD
北大核心
1999年第4期340-344,共5页
Acta Geodaetica et Cartographica Sinica
