I use forecast package in R.
Hyndman says:
The arima() function in R (and Arima() and auto.arima() from the forecast package) fits a regression with ARIMA errors.
I have an output for auto.arima()
Regression with ARIMA(5,0,0) errors
Coefficients:
ar1 ar2 ar3 ar4 ar5 intercept xreg1 xreg2 xreg3 xreg4
xreg5 xreg6 xreg7 xreg8 xreg9
0.0212 -0.0530 0.7005 -0.0232 0.0334 862.0474 -4e-04 -0.0303 -0.0659 -0.1657 4.4673 0.1958 0.3381 -0.4270 5.3478
s.e. 0.0087 0.0086 0.0062 0.0087 0.0087 285.6206 1e-04 0.0604 0.0648 1.7225 0.5952 0.0213 0.0138 0.1415 0.0707
sigma^2 = 15.05: log likelihood = -37334.05
AIC=74700.1 AICc=74700.14 BIC=74820.22
Training set error measures:
ME RMSE MAE MPE MAPE MASE ACF1
Training set -0.0001219744 3.877156 1.434961 NaN Inf 0.3321699 -0.007453887
Can I rename all xreg variables somehow and have real names in my summary output?
Name the columns of the matrix to whatever you like.
library(forecast)
xreg <- ts(matrix(rnorm(900), ncol=9))
colnames(xreg) <- LETTERS[1:9]
auto.arima(WWWusage, xreg=xreg)
#> Series: WWWusage
#> Regression with ARIMA(1,1,1) errors
#>
#> Coefficients:
#> ar1 ma1 A B C D E F G
#> 0.6418 0.6653 0.1122 0.2939 0.0958 0.0923 -0.3412 0.0706 -0.0008
#> s.e. 0.0842 0.0927 0.0986 0.1038 0.0904 0.1061 0.1218 0.0919 0.1074
#> H I
#> 0.0137 -0.2185
#> s.e. 0.0773 0.1320
#>
#> sigma^2 = 9.524: log likelihood = -247.12
#> AIC=518.24 AICc=521.87 BIC=549.38
Created on 2022-03-01 by the reprex package (v2.0.1)