(1-B)^d y = X beta + n with phi(B) n = theta(B) e. Stage one is OLS of the differenced response on the differenced covariates; stage two is the Hannan-Rissanen two-stage regression on the residual series.
References
Box and Jenkins (1976), Chapters 10-11 (transfer-function models); disturbance estimation by Hannan and Rissanen (1982), Biometrika 69(1), 81-94. Standard published form; neither source is in the local corpus and neither was read.
Examples
V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Arimaxhr(V, V)
#> $beta
#> [1] 1
#>
#> $phi
#> [1] 0
#>
#> $theta
#> [1] 0
#>
#> $intercept
#> [1] 6.097555e-16
#>
#> $sigma2
#> [1] 7.888609e-31
#>
#> $noise
#> [1] -2.220446e-16 0.000000e+00 4.440892e-16 8.881784e-16 8.881784e-16
#> [6] 8.881784e-16 1.776357e-15 1.776357e-15
#>
#> $resid
#> [1] -4.440892e-16 -4.440892e-16 4.440892e-16 4.440892e-16
#>
#> $p
#> [1] 1
#>
#> $q
#> [1] 1
#>
#> $d
#> [1] 0
#>
#> $r
#> [1] 1
#>
#> $nobs
#> [1] 4
#>
#> $method
#> [1] "ARIMAX by OLS plus Hannan-Rissanen errors (Box-Jenkins 1976)"
#>
#> attr(,"class")
#> [1] "morie_rich_result" "list"
