04-22-2013 03:46 PM
How is this done? I'd like to use the ODS output to exhibit a simple forecast from a polynomial regression. Id like both axis minimums to be 0, but the ODS output fits it to the data.
04-22-2013 03:59 PM
You have to find the template and modify it.
It isn't as scary as it seems at first .
EDIT: I should say, one way...there are possibly others.
04-22-2013 04:15 PM
Thanks Reeze, I'm looking through that and not really seeing how to get it done. Though it would seem the ODS output is not what im looking for, since it only seems to plot a single X against the Y.
The code for the regression I'm trying to plot is as follows:
proc reg data=revreg2;
model revenue=xr2 XR /white spec;
So basically I want my Y to be revenue and my X to be xr2 + XR. What would be the best way to do this graphically, with confidence bands? And with being able to control the max and min on the axes?
04-22-2013 04:24 PM
If you can extract the data from proc reg using an output or out dataset you can plot it manually using PROC SGPLOT.
Then you'll have full control. Not sure where/how you'll get the confidence bands though.
04-22-2013 04:40 PM
Set ODS TRACE ON and find the template REG procedure to create the graph. Then, in the template find the LAYOUT OVERLAY statement. Set the following option.
LAYOUT OVERLAY / XAXISOPTS=(LINEAROPTS=(VIEWMIN=0));
If using Reeza's suggestion, use SGPLOT procedure to render the plot. Use XAXIS MIN=0 to get the right axis scaling.