data a;
   do i=1 to 100;
      x1=rannor(1);
      x2=rannor(2);
      x12=x1*x2;
      y=3+x1+x1*x2+rannor(3);
      output;
   end;
run;
proc reg data=a;
   model y=x1 x12;
quit;
自己另外生成一个变量
令该变量等于x1*x2即可

楼上正解，SAS帮助文档中有具体例子
#You can address this lack of fit by using polynomials of degree 2 for these two variables as shown in the following statements: #
data baseball;
   set sashelp.baseball(where=(name^="Rose, Pete"));
   YrMajor2 = yrmajor*yrmajor;
   CrHits2  = crhits*crhits;
run;

proc reg data=baseball
      plots=(diagnostics(stats=none) RStudentByLeverage(label)
             CooksD(label) Residuals(smooth)
             DFFITS(label) DFBETAS ObservedByPredicted(label));
   id name team league;
   model logSalary = nhits nruns nrbi nbb yrmajor crhits
                     yrmajor2 crhits2;
run;

您好，构建x1*x2时，直接在SAS里做乘法运算吗？那么x1和x2的编码不能等于0吧？还是构建n*m-1项的哑变量？