topic Re: Interpreting Proc ARIMAX output in SAS Academy for Data Science

Interpreting Proc ARIMAX output

WWD — Tue, 26 Feb 2019 22:06:19 GMT

In Module 3, there is a demonstration titled “Demo: Estimation, Residual Analysis and Goodness of Fit”. At the very end of the video, there are two estimates produced by SAS:

Model for Variable kW_Gen

0.520193

Autoregressive Factors

Factor 1 = 1-0.708913B**(1)

I can find earlier in the output the estimate of mu. In that same area, the estimate for the AR1,1 term is 0.708913.

Earlier in the course, AR(1) is:

Y_t = mu(1-phi_1) + phi_1*Y_(t-1) + e_t

My question is what does Factor 1 represent, in general, and what is “B**(1)” in particular?

Thank you.

Re: Interpreting Proc ARIMAX output

Cynthia_sas — Tue, 26 Feb 2019 23:34:46 GMT

Hi:
Thanks for clarifying your question. The instructors who monitor this forum will take a look and respond.
Cynthia

Re: Interpreting Proc ARIMAX output

Cynthia_sas — Wed, 27 Feb 2019 14:31:32 GMT

Hi:

In this demo, the Autoregressive Factor is shown at the end of the output.

The instructor for this class said that

. "The Autoregressive factors writes the AR part of the model in ‘factored form’. You can think of multiplying the (1 – 0.7B**(1)) by y to yield the AR part of the model: y(t) – 0.7* y(t-1). In other words, the B is a backshift operator."

Hope this helps explain the notation in the output.

Cynthia