Help using Base SAS procedures

Forecasting using proc panel results

Reply
Learner
Posts: 1

Forecasting using proc panel results

Hi-

I would like to forecast using results from my panel regressions. I am running either a one-way fixed effects or one-way random effects model (see code below). The unobserved effect that I am considering is an electric utility (i.e., utility_ID).

More specifically, I would like to make the forecasts for each utility. In order to this, however, I will need the coefficients from the model results and the intercepts for each utility ID. I have the model coefficients, but I cannot figure out how to output the intercepts for each utility (fixed or random effect).

I have two questions:

(1) How do I output the modeled intercepts for each utility (i.e, the unobserved effects) for both the fixed and random effects models?

(2) If the random effects model does not produce intercepts, then how can I forecast my results down to the utility-level?

I look forward to hearing from you. Please feel free to suggest alternative code to accomplish this.

Thanks,

Pete

-------

%macro tests(regr);

ods results off;
ods listing close;

ODS TAGSETS.msoffice2k
file="C:\Users\phl\Desktop\Interruption Analysis\Reliability Model\Regressions\test_&regr..xls"
STYLE=journal
OPTIONS ( Orientation = 'landscape' FitToPage = 'yes' Pages_FitWidth = '1' Pages_FitHeight = '100' );
goptions reset = all;

proc panel data=relia.test outest=test_output_&regr.;
id utility_id year;
model SAIFI = Total_Sales_MWh_Cust Wind_POS_DEV Wind_POS_sqr_DEV Year / &regr. vcomp=wk plot=fitplot HCCME=1 bp STATIONARITY(ALL);
run;

ods tagsets.msoffice2k close;
%mend;
%tests (fixone);
%tests (ranone);
Ask a Question
Discussion stats
  • 0 replies
  • 301 views
  • 0 likes
  • 1 in conversation