如何通过MA模型提前30天预测?

时间:2016-07-13 12:35:12

标签: r

我刚刚从包arma() tseries估算了一个MA模型,其中包含第5个第9个和第14个成分:

Coefficient(s):
       ma5         ma9        ma14   intercept  
-0.0384602  -0.0543772   0.0973954   0.0002656

但现在不要提前30天预测。

1 个答案:

答案 0 :(得分:0)

您可以使用ma参数进行设置。

这是一个最小的,完整的,可验证的例子:

require(tseries)
data(nino)
s <- nino3.4
summary(s.arma
        <- arma(s, lag=list(ar=c(1,3,7,10,12,13,16,17,19),ma=30)))
Call:
arma(x = s, lag = list(ar = c(1, 3, 7, 10, 12, 13, 16, 17, 19),     ma = 30))

Model:
ARMA(19,30)

Residuals:
      Min        1Q    Median        3Q       Max 
-1.374330 -0.225439  0.003465  0.220010  1.087211 

Coefficient(s):
           Estimate  Std. Error  t value Pr(>|t|)    
ar1         1.08886     0.02627   41.447  < 2e-16 ***
ar3        -0.18211     0.03013   -6.044 1.50e-09 ***
ar7        -0.13282     0.02269   -5.854 4.79e-09 ***
ar10        0.17785     0.02934    6.062 1.35e-09 ***
ar12        0.16529     0.04882    3.386  0.00071 ***
ar13       -0.26895     0.04282   -6.281 3.37e-10 ***
ar16       -0.22409     0.04418   -5.073 3.92e-07 ***
ar17        0.25495     0.04731    5.389 7.07e-08 ***
ar19       -0.04727     0.02619   -1.805  0.07110 .  
ma30        0.04283     0.04141    1.034  0.30095    
intercept   4.53718     0.76437    5.936 2.92e-09 ***

有用的资源:

https://cran.r-project.org/web/views/TimeSeries.html

http://a-little-book-of-r-for-time-series.readthedocs.io/en/latest/

https://www.otexts.org/fpp

https://cran.r-project.org/doc/contrib/Ricci-refcard-ts.pdf