topic Re: Looking for simple codes to put mutiple coefficients in SAS Programming

Looking for simple codes to put mutiple coefficients

JKCho — Fri, 19 Jul 2019 19:59:53 GMT

proc surveyreg data=crdsimple;
class sic fyear idvc ids loss;
model rating = sic fyear logat roa lev idvc ids loss/ solution;
run;
data crd2;
set crdsimple;

rtmade = 16.8313718 -0.9310491*logat-1.0102761*roa+0.7764174*lev-2.4251711*idvc-0.4183952*ids+1.5549358*loss ;
run;

data crd3;
set crd2;
if sic = 0100 then rtmade1 = rtmade+2.9293598;
if fyear = 1993 then rtmade2 = rtmade1-3.5688886;
if sic = 1000 then rtmade1 = rtmade+2.4553097;
if fyear = 2008 then rtmade2 = rtmade1--1.0594573;
run;

Hi!!!!

Throughout the above codes, I could obtain coefficient to estimate by dependent variable!

However, the problem is... that I have so many categorical variables and need to input so many coefficients of SICs and FYEARs.

As you can see the last code, I did a little but I have more than 500 categories...

Does anyone know better ways to code to reflect categorical coefficients?

I appreciate your helpful comments so much!!!!

Re: Looking for simple codes to put mutiple coefficients

ballardw — Fri, 19 Jul 2019 20:50:09 GMT

@JKCho wrote:
proc surveyreg data=crdsimple;
class sic fyear idvc ids loss;
model rating = sic fyear logat roa lev idvc ids loss/ solution;
run;
data crd2;
set crdsimple;
rtmade = 16.8313718 -0.9310491*logat-1.0102761*roa+0.7764174*lev-2.4251711*idvc-0.4183952*ids+1.5549358*loss ;
run;
data crd3;
set crd2;
if sic = 0100 then rtmade1 = rtmade+2.9293598;
if fyear = 1993 then rtmade2 = rtmade1-3.5688886;
if sic = 1000 then rtmade1 = rtmade+2.4553097;
if fyear = 2008 then rtmade2 = rtmade1--1.0594573;
run;
Hi!!!!

Throughout the above codes, I could obtain coefficient to estimate by dependent variable!

However, the problem is... that I have so many categorical variables and need to input so many coefficients of SICs and FYEARs.

As you can see the last code, I did a little but I have more than 500 categories...

Does anyone know better ways to code to reflect categorical coefficients?

I appreciate your helpful comments so much!!!!

So, where are those coefficients coming from? If you mean that you need to use something from the output of the Surveyreg procedure then you should direct the needed output into a data set from the proc.

maybe

proc surveyreg data=crdsimple;
class sic fyear idvc ids loss;
model rating = sic fyear logat roa lev idvc ids loss/ solution;
ods output Parameterestimates =work.parms;
run;

And work.parms would have the coefficients needed.

You would likely need to transpose modify the output set to use in the fashion you are attempting.

Re: Looking for simple codes to put mutiple coefficients

Reeza — Sat, 20 Jul 2019 00:13:26 GMT

If you're looking to take a model and apply those results to new data, the term is often called Scoring a new data set and there's PROC SCORE, PROC PLM or SCORE statements within procs that automate this process so you don't need to code it yourself.

Re: Looking for simple codes to put mutiple coefficients

JKCho — Sat, 20 Jul 2019 05:46:40 GMT

A quick solution was simple out p=pred... but I learned stm new from you too! Thx!!!