260. VARIMAX

Vector ARIMA with exogenous regressors

= endogenous vector series
= exogenous regressor vector
= matrix of regression coefficients
Each of the series can depend on the same set of exogenous variables.

Parameters: , , (), ,
Orders: , , , , (regressors)

Example:

Given

  • Orders: , , , , (one exog)
  • Coefficient matrices:

  • Intercept:
  • Initial conditions: ,
  • Endogenous :
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
12 10 8 11 14 12 9 13 16 14 11 15 18 16 13 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
8 9 7 6 10 11 9 8 12 13 11 10 14 15 13 12
  • Exogenous :
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Step 1 — formula

Substitute orders into the VARIMAX recursion (with , no differencing):

Forecast (set ):

Three contributions: AR product, exogenous regression , MA product.

Componentwise:

Innovation:

Step 2 — apply at

Compute the three contributions separately, then sum.

AR part:

Exog part:

MA part: . With , the step gives and .

Sum:

Step 3 — iterate

Three contributions per row: AR , exog , MA . Values rounded to 4 decimal places.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16