How to run models based on first N variables from a list

lionking19063 — Fri, 04 Nov 2022 14:13:02 GMT

Hi,

I wonder how to run models based on first N input variables from a list. For example, the list contains "age height weight BMI income" and I would like to fit a series of regression model based on first N variable. I can write them one by one, but wonder if there is a efficient way of doing it. Thank you in advanced.

model 1:
proc reg data=data;

model y=age;

run;

model 2:
proc reg data=data;

model y=age height;

run;

model 3:
proc reg data=data;

model y=age height weight;

run;

model 4:

proc reg data=data;

model y=age height weight BMI;

run;

model 5:

proc reg data=data;

model y=age height weight BMI income;

run;

Re: How to run models based on first N variables from a list

ballardw — Fri, 04 Nov 2022 14:47:25 GMT

Where does your list reside?

With Proc Reg I think you are looking at macro language. Perhaps something like:

%macro reglist(varlist= );
   %let words=;
   proc reg data=data;

   %do i= 1 %to %sysfunc(countw(&varlist));
      %let words= &words %scan(&varlist,&i.);
      model&1.: model y=&words.;
      /*optional Model&i statements here*/
   %end;
   run;
%mend;

%reglist (varlist=age height weight BMI income);

Proc Reg will allow multiple model statements. If you are adding other optional statements then all of them related to a single model would have to appear before the next model statement. The bit before the model=, where I am using Model&i is a label that will have the text Model1, Model2, etc, depending on the number of variables used appearing in the output so you can tell which model the output is for.

topic How to run models based on first N variables from a list in SAS Programming

How to run models based on first N variables from a list

Re: How to run models based on first N variables from a list