GM(1,1)-ARMA(n,m)预测模型及应用

GM(1,1)模型是以灰色系统理论为基础的预测模型,而ARMA(n,m)是以时间序列为基础的预测模型.由于二模型各具其特点,并各有局限性.但二者能取长补短.文献[1]给出了混合模型GM(1,1)-ARMA(n,m),但在具体预测时由于运算量大,没有合适的计算机处理程序使之显得不方便.本文根据其原理,完善了计算方法,给出了预测公式.经过例子验证,结果是理想的.

The calculating pattern and application of GM(1,1)-ARMA(n,m)

: The model of GM(1,1) is based on the grey system , while the pattern of ARMA(n,m)is based on the theory time series nanalysis . For the above two models , it is becaus of their respective shortcoming and advantage that reference views a mixed model GM(1,1)-ARMA(n,m) , particularly,the pattern of GM(1,1)-AR(n ,m) . However , we feel it is difficult to use it because of its complex caculating . This paper improves its caculating method and shows prediction formula . In short , not only does this paper enlarge the theory of prediction , but the method and computer programme are feasible through two examples test .