## Need help in determining the autoregressive model equation and predictions

Occasional Contributor
Posts: 6

# Need help in determining the autoregressive model equation and predictions

I have developed a model using regression analysis with autoregressive errors, using the maximum likelihood method.

I am able to figure out the model equation but when i try to put everything in an equation I get different results.

 Parameter Estimates Variable DF Estimate Standard t Value Approx Error Pr > |t| Intercept 1 92.5242 6.2288 14.85 <.0001 T 1 -1.2082 0.0741 -16.3 <.0001 DP 1 -0.6514 0.1051 -6.2 <.0001 T*DP 1 0.0171 0.00111 15.37 <.0001 DP*H 1 0.8459 0.0431 19.61 <.0001 H 1 -56.4553 3.705 -15.24 <.0001 AR1 1 -0.989 0.001651 -599.11 <.0001

The MSE for this analysis is 16.03

The equation would be

y=92.5242-1.2T-0.65DP+0.017T*DP+0.84DP*H-56.45H+Vt

Vt=0.989V(t-1)+16.03

How do I calculate the V(t-1)? as it varies for each observation.

Posts: 5,529

## Re: Need help in determining the autoregressive model equation and predictions

Look at the documentation for the AUTOREG procedure at SAS/ETS(R) 9.3 User's Guide.

The recursive expression is given in the section "Predicting Future Series Realizations".

PG

PG
Occasional Contributor
Posts: 6

## Re: Need help in determining the autoregressive model equation and predictions

I tried. but its not working.

Discussion stats
• 2 replies
• 179 views
• 3 likes
• 2 in conversation